Method of calibrating a multi-spectral vision system

Abstract

A method, system, and non-transitory CRM of calibrating a vision system includes a perforated calibration board and an optical device. Calibrating the vision system further includes positioning at least one of the optical devices or the perforated calibration board at a plurality of positions relative to each other; capturing, via the optical device, a plurality of images of the perforated calibration board at those positions; processing the images to analyze the perforated calibration board; establishing a correction by estimating parameters of the optical device based on the analysis; and applying the correction to subsequent images captured by the optical device.

Description

PATENT33894 / H011METHOD OF CALIBRATING A MULTI-SPECTRAL VISION SYSTEM Cross-Reference to Related Applications

[0001] The present application claims the priority benefit of U.S. Provisional Application No. 63 / 827,621, filed June 20, 2025, U.S. Provisional Application No. 63 / 742,694, filed January 7, 2025, and U.S. Provisional Application No. 63 / 672,691, filed July 17, 2024, the entire disclosures of which are hereby incorporated by reference.Field of the Disclosure

[0001] The present disclosure generally relates to vision systems and more specifically to a method for calibrating a multi- spectral vision system and using the multi-spectral vision system with a fiducial.Background of the Disclosure

[0002] Existing optical calibration techniques typically involve a precisely referenced target that can be identified in sensor imagery. These techniques work for visible light and some other spectra but have difficulty with long-wave infrared (LWIR) light. LWIR calibration techniques typically involve high contrast materials which are expensive and hard to come by. Further, LWIR calibration techniques typically involve a collimator, special illumination sources, and devices for extremely precise rotation of the sensor. Resultingly, calibration setups for long wave-infrared (LWIR) cost between $50,000 to $100,000 for the base equipment alone. Moreover, existing calibration techniques make it difficult to position LWIR sensors relative to sensors of a different spectrum, or even relative to each other with enough precision to do calibration (e.g., for binocular stereo).Summary

[0003] In some aspects, the techniques described herein relate to a method of calibrating a vision system, including: positioning at least one of (i) an optical device or (ii) a perforated calibration board at a plurality of positions relative to the other of (i) the optical device or (ii) the perforated calibration board; capturing, via the optical device, a plurality of images of the perforated calibration board at the plurality of positions; processing the plurality of images to analyze the perforated calibration board; establishing a correction by estimating parameters of the optical device based on the analyzed perforated calibration board; and applying the correction to subsequent images captured by the optical device.

[0004] In some aspects, the techniques described herein relate to a system for calibrating a vision system, including: a perforated calibration board; an optical device of the vision system;PATENT33894 / H011 one or more processors; and one or more memories storing computer-executable instructions that cause the one or more processors to at least: position at least one of (i) the optical devices or (ii) the perforated calibration board at a plurality of positions relative to the other of (i) the optical device or (ii) the perforated calibration board; capture, via the optical device, a plurality of images; process the images to analyze the perforated calibration board; establish a correction by estimating parameters of the optical device based on the analyzed perforated calibration board; and apply the correction to subsequent images captured by the optical device.

[0005] In some aspects, the techniques described herein relate to one or more non-transitory computer-readable medium having computer-executable instructions stored thereon that, when executed by one or more processors, cause the one or more processors to at least: position at least one of (i) an optical device or (ii) a perforated calibration board at a plurality of positions relative to the other of (i) the optical device or (ii) the perforated calibration board; capture, via the optical device, a plurality of images; process the images to analyze the perforated calibration board; establish a correction by estimating parameters of the optical device based on the analyzed perforated calibration board; and apply the correction to subsequent images captured by the optical device.Brief Description of the Drawings

[0006] The features of this invention which are believed to be novel arc set forth with particularity in the appended claims. The invention may be best understood by reference to the following description taken in conjunction with the accompanying drawings, in which like reference numerals identify like elements in the several FIGS., in which:

[0007] FIG. 1A illustrates an example flowchart of a method of calibrating a multi- spectral vision system in accordance with the principles of the present disclosure.

[0008] FIG. IB illustrates two examples of perforated calibration board for use in the method of FIG. 1 A, one made of PVC (left) and the other made of aluminum (right).

[0009] FIG. 1C is a close-up of one of the perforated calibration boards of FIG. IB.

[0010] FIG. ID illustrates a perspective view of another example of a perforated calibration board for use in the method of FIG. 1 A.

[0011] FIG. IE illustrates a cross-sectional view of bevels that can be incorporated into the perforated calibration boards of FIG. IB.

[0012] FIG. IF illustrates an example image of a perforated calibration board positioned relative to a background.PATENT33894 / H011

[0013] FIG. 1 G illustrates an example image of the perforated calibration board positioned relative to the background of FIG. IF.

[0014] FIG. 1H illustrates another example image of the perforated calibration board positioned relative to the background of FIG. IF.

[0015] FIG. II illustrates example images of unheated perforated calibration boards (left column) and heated perforated calibration boards (right column) taken from various distances.

[0016] FIG. 1 J illustrates an example image of the perforated calibration boards of FIG. IB when unheated.

[0017] FIG. IK illustrates an example image of the perforated calibration boards of FIG. IB when heated.

[0018] FIG. IL illustrates an example image of an UV print-on-PVC target showing nominal contrast and dot pattern visible.

[0019] FIG. IM illustrates an example image of the UV print-on-PVC target of FIG. IL with illumination from the front (left) and from the back (right) with no dot pattern visible.

[0020] FIG. IN illustrates an example image of a cardboard box with a hole in the front and the back intact from less than 5 meters away.

[0021] FIG. 10 illustrates an example image of a heated aluminum perforated calibration board with diffuse reflection of heat lamp.

[0022] FIG. IP illustrates an example image of an unheated aluminum perforated calibration board with reflection of overhead lights.

[0023] FIG. IQ illustrates an example image of unheated perforated calibration boards of FIG. IB taken from ~l-2m away.

[0024] FIG. 2A is a front, perspective view of one example of a multi- spectral vision system that can be calibrated using the method of FIG. 1 A.

[0025] FIG. 2B is another front, perspective view of the vision system of FIG. 2A.

[0026] FIG. 2C is a front, perspective view of a housing of the vision system of FIG. 2A.

[0027] FIG. 2D is an exploded view of the housing of FIG. 2C.

[0028] FIG. 2E is another front, perspective view of the vision system of FIG. 2A.

[0029] FIG. 2F is a rear, perspective view of a camera module and an electronics plate of the vision system of FIG. 2A.

[0030] FIG. 2G is a front, exploded view of the camera module.

[0031] FIG. 2H is a rear, exploded view of the camera module.PATENT33894 / H011

[0032] FIG. 21 is a front, perspective view of a lens cover of the camera module.

[0033] FIG. 2J is a rear view of FIG. 21.

[0034] FIG. 2K is a top view of FIG. 21.

[0035] FIG. 2L is a cross-sectional view taken along line C-C in FIG. 2K.

[0036] FIG. 2M is a perspective view of an electro-optical camera of the vision system of FIG. 2A.

[0037] FIG. 2N is a perspective view of a bracket for the electro-optical camera of FIG. 2M.

[0038] FIG. 20 is a perspective view of a bracket for infrared cameras of the vision system of FIG. 2A.

[0039] FIG. 2P is a front, perspective view of the electronics plate of the vision system of FIG. 2A.

[0040] FIG. 2Q is a rear, perspective view of the electronics plate.

[0041] FIG. 2R illustrates the electronics plate being coupled to the housing.

[0042] FIG. 2S is a cross-sectional view of the vision system of FIG. 2A.

[0043] FIG. 2T is an exploded view of the vision system of FIG. 2A.

[0044] FIG. 2U is a front, perspective view showing the orientation of the vision system of FIG. 2A.

[0045] FIG. 2V is a side view of FIG. 2U.

[0046] FIG. 2W is a real', perspective view of FIG. 2U.

[0047] FIG. 3A illustrates an example computing environment for calibrating a vision system using the method of FIG. 1A.

[0048] FIG. 3B illustrates an example Python script for performing a comer detection estimate.

[0049] FIG. 3C illustrates an example image with the corner detection estimate overlayed.

[0050] FIG. 3D illustrates an example image with a checkerboard pattern.

[0051] FIG. 4A is a first perspective view of an example of a calibration object in the form of an IR fiducial that is constructed in accordance with the principles of the present disclosure and can be used in the calibration method of FIG. 1 A.

[0052] FIG. 4B is a second perspective view of the IR fiducial of FIG. 4A.

[0053] FIGS. 4C and 4D illustrate example images of the IR fiducial of FIG. 4A after being heated for a short period of time.

[0054] FIGS. 4E and 4F illustrate example images of the IR fiducial of FIG. 4A after being heated relative to people.PATENT33894 / H011

[0055] FIGS. 4G-4K illustrate example images of another example of an IR fiducial from a variety of distances and angles, the IR fiducial having circular perforations.

[0056] FIGS. 4L-4Q illustrate example images of the IR fiducial of FIGS. 4G-4K in stereo.

[0057] FIG. 4R illustrates an example image of the IR fiducial of FIGS. 4G-4K with shadows.

[0058] FIGS. 4S and 4T illustrate example images of the IR fiducial of FIGS. 4G-4K with fasteners showing and edge effects.

[0059] FIG. 5A illustrates an example image of an IR fiducial with mounting.

[0060] FIG. 5B illustrates an example isometric schematic view of an IR fiducial with mounting.

[0061] FIG. 5C illustrates an example isometric schematic view of the rear of an IR fiducial with mounting.

[0062] FIG. 5D illustrates an example isometric schematic view of the front of a FSR calibration object.

[0063] FIG. 5E illustrates an example isometric schematic view of the rear of the FSR calibration object of FIG. 5D.

[0064] FIG. 5F illustrates an example image of an IR fiducial.

[0065] FIG. 5G illustrates an example image and schematic view of the rear of an IR fiducial with a thermal control block diagram.

[0066] FIG. 5H illustrates another example isometric schematic view of an FSR calibration object.

[0067] FIG. 51 illustrates another example isometric schematic view of the rear of the FSR calibration object of FIG. 5H.

[0068] FIG. 5J illustrates an example schematic view of the rear of the FSR calibration object of FIG. 5H.

[0069] FIG. 5K illustrates an example schematic view of the rear of the FSR calibration object of FIG. 5D.

[0070] FIG. 6 illustrates an example system for image calibration confirmation.Detailed Description

[0071] The present disclosure aims to reduce problems with conventional techniques (e.g., as described in the Background section) by providing techniques for calibrating a multi-spectral vision system.PATENT33894 / H011

[0072] Advantageously, the present techniques image processing methods (e.g., adaptive thresholding and iterative, gradient -based sub-pixel refinement) using parallel processing improve calibration computational speed, precision, and real-time accuracy. The present techniques enable calibration feature extraction, correction parameter computation, and sensor alignment with minimal error even in complex, low-contrast, or thermally challenging environments. As a result, the present techniques enable a dramatic improvement in the alignment of multi-spectral sensors for tasks like 3D reconstruction and pose estimation, ensuring that calibration errors are minimized.

[0073] Advantageously, a dual-board fiducial configuration provides a stable thermal benchmark with thermal and structural stability that is maintainable over a range of operating conditions. One board may be actively heated while being thermally isolated from the other, with precisely engineered bevels and distinct perforation patterns. This design minimizes thermal drift (which would otherwise degrade calibration accuracy) and establishes consistent, high-contrast reference features that enhance optical detection. The synchronization of thermal states with image capture ensures that temperature-dependent variations are accounted for in the calibration parameters, thereby reducing noise and error.

[0074] Advantageously, an array of cameras with overlapping fields of view and complementary spectral sensitivities (e.g., NIR, SWIR, MWIR, LWIR, and EO) enables multi- spectral calibration verification. By detecting a predetermined calibration board pattern and comparing it against a desired pattern (e.g., user selected), the present techniques calculate deviations to dynamically refine sensor parameters. Real-time network communication among components facilitates continuous sensor fusion and cross-verification, yielding a self-correcting calibration process that enhances system reliability and resilience under variable environmental conditions.

[0075] Additional advantages of the present techniques over conventional approaches will be appreciated throughout this disclosure by one of ordinary skill in the art. Some embodiments may include some, none, or all of the enumerated advantages. The various concepts and techniques introduced above and discussed in further detail below may be implemented in numerous ways, and the described techniques are not limited to any particular manner of implementation. Examples of implementations are provided below for illustrative purposes.

[0076] The present disclosure is generally directed to a method that addresses the problems discussed above and other problems associated with existing optical multi- spectral calibrationPATENT33894 / H011 techniques. For example, the present disclosure is directed to a method of calibrating a vision system that is more cost-effective, robust, and precise than existing optical calibration techniques. The vision system is generally configured to capture, process, and analyze images obtained by one or more image sensors, one or more optical devices, other elements of the vision system, and other data. The vision system can in turn identify or classify the environment (including objects in that environment) in which the vision system is operating.Example Method

[0077] FIG. 1A depicts a flow chart of one example of a method 100 of calibrating a multi- spectral vision system in accordance with the teachings of the present disclosure. The method 100 in this example is a method of performing a spatial calibration of the multi- spectral vision system (as opposed to performing a radiometric calibration of the vision system, though in other examples, the method 100 may be used for that purpose). In some examples, the vision system can be part of a maritime vehicle (e.g., a boat, watercraft, submarine, or amphibious vehicle) primarily intended for military purposes (e.g., for naval defense, patrolling waters and enforcing laws, reconnaissance, naval exploration, monitoring). In some examples, the vision system can be part of a different type of vehicle (e.g., a car, a truck, a train), a manipulative robot (e.g., surgical robots, drones), or any other device or system that could benefit from a calibrated vision system. The vehicle employing the calibrated vision system can be manned (i.e., operated by an onboard human) or unmanned, and unmanned vehicles can be remotely controlled or autonomous, for example.

[0078] As illustrated in FIG. 1A, the multi- spectral calibration method 100 includes the act of positioning a perforated calibration board 112 at a plurality of positions relative to an optical device of the vision system (block 102). The optical device may be one or more of an electro- optical (EG) camera, a stereo camera, or an infrared (IR) camera (e.g., a near-IR, mid-wave IR, long-wave IR camera), for example. Further details about the optical device are discussed below in connection with FIG. 2A.

[0079] The perforated calibration board 112 is preferably planar but has a thickness I sufficient to provide rigidity. The perforated calibration board 112 is also preferably made of a material that is structurally rigid enough to resist or prevent warping (e.g., when heated). FIG. IB illustrates two exemplary perforated calibration boards 112. And as illustrated in FIG. IB, the perforated calibration board 112 is preferably made of a reflective material such as aluminum (right most perforated calibration board 112 in FIG. IB). However, one of skill in the art wouldPATENT33894 / H011 appreciate that the perforated calibration board 112 can be manufactured of any suitable material, including but not limited to soda-lime glass, PVC (left most perforated calibration board in FIG. IB), or another suitable material.

[0080] The perforated calibration board 112 generally includes a first (or front) side and a second (or rear) side, and one or more perimeter edges 114 that extend between and connect the front side and the rear side and, as such, define the thickness t of the board 112. The perforated calibration board 112 also generally includes one or more perforations 116 (which can also be referred to as holes or apertures) located in / within the solid area defined between the one or more perimeter edges 114. The perforations 116 are generally arranged in a pre-determined pattern determined by the specific calibration technique, use case, implementation, etc. (e.g., to enable self-similarity and invariance under rotation). For example, the perforations 116 can be arranged in a plurality of alternating rows and columns. The perforations 116 can have any number of different geometric shapes (e.g., a rectangular shape, a square shape, a triangular shape, a circular shape, an oval shape, an elliptical shape), non-geometric shapes, or any suitable combinations thereof. For example, each of the perforated calibration boards 112 illustrated in FIG. IB has a rectangular shape defined by four perimeter edges 114 as well as eight square shaped perforations 116 located within the solid area defined by and between the four perimeter edges 116. In other words, each of the square shaped perforations 116 is spaced from the perimeter edges 114, as is best illustrated in FIG. 1C. Each of the perforations 116 is generally defined by one or more interior edges 118 of the perforated calibration board 112 that span the thickness t of the board 112. In the example illustrated in FIG. IB, each of the perforations 116 is defined by four interior edges 118. FIG. ID illustrates another example arrangement of perforated calibration board 112 having four perimeter edges 114 and a plurality of square shaped perforations 116 located within the solid area defined by and between the four perimeter edges 114.

[0081] One or more of the perforations 116 of the perforated calibration board may be defined by one or more bevels 120 of suitable angle and dimensions so that the thickness of the perforated calibration board 112 (which is needed for rigidity) is not visible when positioning the perforated calibration board 112 at a plurality of positions relative to the optical device. By ensuring the thickness of the perforated calibration board 112 is not visible, the optical device may be calibrated more precisely at a greater number of positions (e.g., angles). More particularly, in some examples, one or more of the interior edges 118 that define one or more ofPATENT33894 / H011 the perforations 116 can be beveled relative to the first (or front) side of the perforated calibration board 112. For example, as illustrated in FIG. IE, the interior edges 118 that define one (or more) of the square shaped perforations 116 can be beveled relative to the front side of the perforated calibration board 112. In the example illustrated in FIG. IE, the bevels 120 are formed at an angle of approximately 45 degrees relative to the front side of the perforated calibration board 112. In other examples, however, the bevels 120 can be formed at a different angle relative to the front side of the perforated calibration board 112. Additionally or alternatively, one or more of the perimeter edges 114 of the perforated calibration board 112 can be beveled. In addition to the benefits briefly discussed above, it will be appreciated that beveling the perforations 116 and / or the perimeter edges 114 of the perforated calibration board 112 can also beneficially reduce the overall weight of the perforated calibration board 112 allowing for easier, faster, and more accessible calibration.

[0082] Additionally or alternatively, the interior edges 118 of the perforations 116 and / or the edges 114 of the perforated calibration board 112, may be a different color than the rest of the perforated calibration board 112 so as to make the thickness of the perforations 116 and / or the thickness t of the perforated calibration board 112 effectively invisible to the optical device. In turn, the perimeter edges 114 and / or the interior edges 118 may be associated with the background (e.g., the background material) rather than the foreground (e.g., the perforated calibration board 112) when calibrating the optical device, for example. In some examples, the interior edges 118 of the perforations 116 and / or the edges 114 of the perforated calibration board 112 can be the exact same color (e.g., a dark color like black or the like) as the background. In other examples, the interior edges 118 of the perforations 116 and / or the edges 114 of the perforated calibration board 112 can be a color that more closely matches the color of the background (as compared to the color of the rest of the perforated calibration board 112). For example, when the background has a dark color like brown, the interior edges 118 of the perforations 116 and / or the edges 114 of the perforated calibration board 112 can be a color that is darker (and thus closer to the color of the background). In some examples, the background may be a lighter and / or different color like white and the interior edges 118 of the perforations 116 and / or the edges 114 of the perforated calibration board 112 can be a color that is lighter and / or different (and thus closer to or further from the color of the background).

[0083] Because the edges of the perforated calibration board 112 may be the same color as the background material the edges may only be visible to EO cameras. Configuring the edges to bePATENT33894 / H011 the same color as the hackground material is beneficial because this configuration is cheaper, easier, and more likely to yield precise results than beveling. LWIR cameras may have a lower resolution, therefore obscuring the edges during calibration may not be necessary. The colored edges may not obscure the edge when calibrating an LWIR camera, for example. It is to be understood that the colored edges may be and / or include the bevels 120, thereby enabling greater cross calibration techniques across various detectors (e.g., any IR camera up the spectrum to EO cameras, etc.).

[0084] Additionally or alternatively, the perforated calibration board 112 may be positioned at the plurality of positions relative to a background 122. The background 122 may be made of a cloth material (e.g., an absorbative material with good contrast relative to the perforated calibration board 112) or other suitable material (e.g., spray-painted dry wall, concrete, aluminum, etc.). FIG. IF, 1G, and 1H illustrate example images of the perforated calibration board 112 positioned relative to the background 122. FIG. IF and 1G illustrate an example image taken with an EO camera of the perforated calibration board 112 relative to a dark cloth background 122. FIG 1H illustrates an example image taken by an LWIR camera of the perforated calibration board 112 relative to another dark cloth background 122. Alternatively, the background 122 may be defined by pail of an infrared (IR) fiducial, further details about which will be discussed in greater detail below. For example, the background 122 can be defined by a second or rear calibration board positioned behind or rearward of a first or forward perforated calibration board 112. It should be understood that in some examples the background may be defined by the surrounding environment.

[0085] Additionally or alternatively, the perforated calibration board 112 may include one or more mounts (e.g., brackets, adapters, fixtures, holes, etc.) for coupling (e.g., attaching, affixing, securing, connecting, fitting, etc.) one or more attachments 124 (e.g., handles, tripods, connectors, etc.) to the perforated calibration board 112. In the example of FIGS. IF and 1G, the attachment 124 is a tripod used for positioning the calibration board 112. In the example of FIG. 1H, the tripod is not visible.

[0086] In some examples, the perforated calibration board 112 may be heated. The one or more mounts and / or the one or more attachments 124 may be made of an insulating material (e.g., ceramic, polystyrene, fiberglass, garolite, polyvinyl chloride (PVC ), polyoxymethylene (POM), fiberglass, polyetheretherketone (PEEK), nylon, etc.) for safety, accessibility, and comfort reasons. In some examples, a perforated calibration board 112 may be coupled to aPATENT33894 / H011 second calibration board to form an IR fiducial, as will be discussed in greater detail below. A plurality of perforated calibration boards 112 and / or IR fiducials may be coupled (e.g., rigidly mounted) to each other to form a pattern of perforated calibration boards 112, IR fiducials, and / or any combinations thereof. Each perforated calibration board 112 and / or IR fiducial may be rigidly mounted to each other to form a distinct (e.g., unique) pattern of perforations 116. The plurality of perforated calibration boards 112 and / or IR fiducials mounted to each other may be weighed down or fastened (e.g., bolted) to a fixed point (e.g., the floor). The plurality of perforated calibration boards 112 and / or IR fiducials may be modular (e.g., using one or more mounts, connectors, etc.) or affixed to each other (e.g., welded). Using a plurality of perforated calibration boards 112 and / or IR fiducials may more precisely calibrate the optical device by using a spatial reference to enable a calibration program to perform loop closure.

[0087] Referring back to FIG. 1A, the multi-spectral calibration method 100 also includes capturing, via the optical device, a plurality of images of the perforated calibration board 112 at the plurality of positions (block 104). As illustrated in FIG. II, a plurality of images of the perforated calibration board 112 may be captured, via the optical device (e.g., an LWIR camera), at a plurality of positions (e.g., distances, angles, etc.). The perforated calibration board 112 may be made of PVC (left most target) or aluminum (right most target) and be unheated (left column) or heated (right column). The images illustrated in FIG. II are taken from Im to 8m away. FIGS. 1J and FIG. IK illustrate the example images of FIG. II of the perforated calibration board 112 from 5m away unheated (FIG. II) and heated (FIG. IK). FIGS. IL and IM illustrate images taken by an EO and LWIR camera, respectively. FIG IL illustrates an example image taken by an EO camera of an ultraviolet (UV) print-on-PVC targets with a visible dot pattern. FIG. IM illustrates an example image taken by an LWIR camera of UV print-on-PVC targets with the dot pattern not visible. In the example of FIG. IM the left target is illuminated from the front and the right target is illuminated from the back. FIG. IN illustrates an example image showing the resolution of edges of a cardboard box with a hole in the front and the back intact from less than 5m away. FIG. 10 illustrates an example image of a heated perforated calibration board 112 with diffuse reflection of heat lamp. FIG. IP illustrates an example image of an unheated perforated calibration board 112 showing reflection of overhead fights. FIG. IQ illustrates an example image of unheated PVC (left) and unheated aluminum (right) from approximately l-2m away.PATENT33894 / H011

[0088] The calibration method 100 further includes processing the plurality of images to analyze the perforated calibration board 112 (block 106). The processing and analysis of the perforated calibration board 112 may include geometry detection, which is further discussed in reference to geometry detection module 314a and FIGS. 3B, 3C, and 3D below. In some examples, processing of the plurality of images generally improves the image quality, thereby improving the analysis / extraction of meaningful information from the image. For example, the processing may include color correction, noise adjustment, sharpening, contrast adjustments, histogram equalization, brightness adjustments, exposure adjustments, local tone mapping, etc. It will be understood that the processing may depend on the type of image (e.g., taken via an LWIR camera, taken via an EO camera, etc.), the quality of the image, and the intended use of the image, for example. In some examples, processing may include converting the image to grayscale. In other examples, processing may include correcting for non-uniform heating in thermal images by fitting and subtracting a model of the intensity data, as well as applying a tophat filter.

[0089] The calibration method 100 further includes establishing a correction by estimating parameters of the optical device based on the analysis of the perforated calibration board 112 (block 108). Establishing a correction and estimating parameters of the optical device are further discussed in reference to the parameter estimation module 314b, below. In some examples the established correction may be a matrix describing the internal geometry of the optical device (e.g., focal length, principal point, etc.) and distortion coefficients quantifying the lens distortion (e.g., radial, tangential, etc.). In other examples, the established correction may include the intrinsic parameters and extrinsic parameters of the optical device (e.g., to facilitate 3D reconstruction, pose estimation, etc.)

[0090] The calibration method 100 further includes applying the correction to subsequent images captured by the optical device (block 110). In some examples, the established correction may be used to correct subsequent images captured by the optical device via undistortian and / or rectification.

[0091] The method 100 may further include heating the perforated calibration board 112 and / or a second calibration board (which can be similar to or different from the perforated calibration board 112) to a one or more temperatures (e.g., a temperature differential, a threshold temperature, a predetermined and / or precalculated temperature, etc.) within an operating range of the optical device. The method 100 may also include capturing, via the optical device, thePATENT33894 / H011 plurality of images for each temperature of the one or more temperatures. In other words, the optical device can capture the radiation emitted by the heated perforated calibration board 112 for each temperature of the one or more temperatures. In general, the one or more temperatures may be calculated (e.g., predetermined) based on the materials selected and environmental factors (e.g., ambient temperature, humidity, dew point, etc.). The one or more temperatures may, in some examples, be one stable temperature (e.g., 30°C) and / or a stable temperature range (e.g., 31.5°C-32.5°C.). In some examples, the method 100 may also include calculating the temperature. In other examples, the temperature can be pre-calculated and obtained for use in the method 100.

[0092] The method 100 may further include positioning the perforated calibration board 112 in a shadowbox for the optical device (e.g., a light-tight box with a controlled internal environment for calibrating the optical device). For example, a shadowbox may be used when calibrating the optical device in a controlled environment (e.g., a manufacturing facility, a maintenance facility, a mobile field service van, etc.)

[0093] The method 100 may further include forming the perforated calibration board 112. Forming the calibration board 112 may include forming one or more perforations in the perforated calibration board using a machine. The machine that forms one or more perforations may be or include a waterjet, laser, or milling machine, for example.

[0094] It should be further understood that the method 100 can be entirely performed in a single location (e.g., on site). Alternatively, it should be understood that the method 100, specifically the positioning, capturing, processing, analyzing, establishing, and applying acts, may be performed in various locations (e.g., in the cloud, at a server in another location, on site, etc.) by various actors (e.g., a person physically positioning the perforated calibration boards 112 and a non-transitory computer readable medium causing an optical device to capture a plurality of images, causing a processor to process the images, establish a correction, and apply the correction). For example, the calibrating act can be performed in the lab, the act of capturing subsequent images can be performed in another location, and the act of applying the correction can be performed in yet another location. This disclosure is in no way meant to be limited to a method in which all of the steps are performed all in the same location or by the same actors (e.g., a person may physically position the perforated calibration board, a machine and / or robot may physically position the perforated calibration boards, etc.). It will be appreciated that the method 100 may be a computer implemented method (e.g., using the components of thePATENT33894 / H011 computing environment 300) in some examples. Additionally the method 100 may be performed in any order with fewer or additional operations.Example Vision System

[0095] FIGS. 2A-2U illustrate one example of a vision system 200 that can be calibrated using the method 100 discussed above. In this example, the vision system 200 is a stereoscopic vision unit that includes two independent stereoscopic cameras. The vision system 200 can, for example, be sealingly and securely mounted to an exterior of a vehicle (e.g., a hull cap of a maritime vehicle). Accordingly, the vision system 200 is optimally positioned to capture, process, and analyze data about the environment surrounding the vehicle.

[0096] The vision system 200 generally includes a housing 204, a camera module 208 coupled to and carried by the housing 204, and an electronics plate 209 coupled to and carried by the housing 204. The housing 204 is configured to be mounted to the exterior of the vehicle. The housing 204 is preferably made of fiberglass but can be made of another strong material such that the housing 204 protects the camera module 208 when the vehicle experiences significant shock (e.g., shock values up to 20G), traverses the body of water at high speeds, or is used in dangerous conditions.In some examples, the housing 204 can be purged with nitrogen or dry air for improved performance. Preferably, and as best illustrated in FIGS. 2C, 2D, and 2T, the vision system 200 also includes a first sealing element (e.g., a gasket) 210 and a mount 211 for mounting the first sealing element 210 to the housing 204. In this example, the mount 211 has a substantially rectangular shape and is coupled (e.g., fixedly or removably coupled) to an outer perimeter edge of the housing 204, and the first sealing element 210 has a similar shape as the mount 211 and is removably disposed in a channel formed in the mount 211. When the first sealing element 210 is mounted to the housing 204 via the mount 211, the first sealing element 210 surrounds the interior of the housing 204, such that the first sealing element 210 is arranged to sealingly engage the camera module 208 and the electronics plate 209 and to effect a seal between the housing 204 and the camera module 208 and the electronics plate 209 when the camera module 208 and the electronics plate 209 are coupled to the housing 204. In other words, the first sealing element 210 serves to seal the electrical components within an interior of the housing 204 when the camera module 208 and the electronics plate 209 are coupled to the housing 204.

[0097] The camera module 208 generally includes a frame 212, one or more electrical-optical (“EO”) cameras 216, one or more infrared (“IR”) cameras 220, a plurality of windows (or lenses)PATENT33894 / H011224, and a plurality of window retention plates 228 for the windows 224. In this example, the frame 212 is defined by a face plate 232 and a cap 236 coupled to the face plate 232 via a plurality of fasteners 237 (one of which is illustrated in FIG. 2H) and via adhesive (e.g., applied to the rear surface of the face plate 232). In other examples, the face plate 232 and the cap 236 can be coupled together in a different manner. Preferably, the frame 212 also includes a second sealing element 238 (e.g., a gasket) secured to the cap 236 so as to be disposed between the face plate 232 and the cap 236. In this example, the second sealing element 238 is secured in a groove formed in the front surface of the cap 236. The frame 212 also includes a plurality of openings 240 formed in the face plate 232 and sized to receive the windows 224. In this example, the frame 212 includes four openings 240 divided into two pairs of openings. In other examples, however, the frame 212 can include more or less openings 240. Moreover, while not illustrated herein, it will be appreciated that the camera module 208 also generally includes a plurality of covers configured to selectively cover the plurality of windows 224, respectively. The plurality of overs may also cover the plurality of apertures 244 formed in the window retention plates 228.

[0098] In this example, the camera module 208 includes a single electrical-optical camera 216 that preferably takes the form of a stereo camera with dual EO image sensors. In this example, the camera module 208 includes two infrared cameras 220. Each of the infrared cameras 220 preferably takes the form of a stereovision IR camera, which may utilize one or more of shortwave IR (SWIR), mid-wave IR (MWIR) (cooled or uncooled), and / or long-wave IR (LWIR) (cooled or uncooled) sensors. When the camera module 208 includes pairs of EO / IR cameras 216, 220, each camera may be configured to capture similar electromagnetic radiation across a similar FOV, and may be separated (e.g., fixedly separated) by a baseline distance.

[0099] In this example, because the frame 212 includes four openings 240, the camera module 208 includes four windows 224. In this example, each window of the plurality of windows 224 is flat, and each window of the plurality of windows 224 is oriented planar to the EO camera 216 and the IR cameras 220. In other examples, however, one or more of the plurality of windows 224 can be oriented non-planar to the EO camera 216 and / or the IR cameras 220. In this example, each of the windows 224 has an anti-reflection coating. Preferably, the windows positioned in front of the EO camera 216 are formed of a substance that is translucent to visible light, and the windows positioned in front of the IR cameras 220 are formed of a substance that is transparent to one or more IR wavelengths. For example, at least two windows of the pluralityPATENT33894 / H011 of windows 224 may be germanium lenses (e.g., manufactured by Edmund Optics). Tn this example, the camera module 208 includes two window retention plates 228, one window retention plate 228 for each of the pairs of openings 240. Thus, in this example, each of the two retention plates 228 has a pair of apertures 244, each aperture 244 sized and arranged to be aligned with a corresponding one of the openings 240 and a corresponding one of the windows 224 when the two retention plates 228 are coupled to the frame 212. In this example, each of the two retention plates 228 is coupled to the frame 212 by disposing each of the retention plates 228 in one of the mounting cavities 246 formed in the frame 212 and inserting a plurality of fasteners 247 (only one of which is shown in FIG. 2G) through the frame 212 and the respective retention plate 228.

[0100] In this example, and as best illustrated in FIGS. 21- 2L, the frame 212 has an outer surface that is curved, creating the appearance that the windows 224 are curved as well (even though they are flat). In this example, each window retention plate 228 has an outer surface that is also curved and is flush with the outer surface of the frame 212. Further, it will be appreciated that the camera module 208 may also include a plurality of first retaining rings 248 and a plurality of second retaining rings 249. The first retaining rings 248, which in this example take the form of shock-resistant threaded retention rings (e.g., manufactured by Thorlabs, inc.), are seated in a channel formed in the frame 212 at a position surrounding the openings 240, respectively. In turn, the first retaining rings 248 are disposed so as to sealingly engage and retain the rear surface of the windows 224, respectively. Meanwhile, the second retaining rings 249 in this example also take the form of shock-resistant threaded retention rings but are seated in a channel formed in one of the window retention plates 228 at a position surrounding a corresponding one of the apertures 244. In turn, the second retaining rings 249 are disposed so as to sealingly engage and retain the front surface of the windows 224, respectively. Beneficially, the first and second retaining rings 248, 249 are configured to use in connection with windows 224 of different thicknesses, such that windows 224 of different thicknesses can be employed in the vision system 200 as needed.

[0101] The camera module 208 also generally includes an EO bracket 250 and an IR bracket 254. The EO bracket 250 is coupled to the frame 212 and is configured to retain the EO camera 216 in position immediately adjacent two of the openings 240. The IR bracket 254 is also coupled to the frame 212 but is configured to retain the IR cameras 220 in position immediately adjacent the other two openings 240.PATENT33894 / H011

[0102] The vision system 200 also includes the electronics plate 209, which is coupled to both the housing 204 and the camera module 208. In this example, the electronics plate 209 is removably coupled to the housing 204 via a plurality of latches 262. In other examples, however, the electronics plate 260 can be removably coupled to the housing 204 in a different manner or can be permanently coupled (e.g., welded) to the housing 204. The electronics plate 260 includes various electrical components for the camera module 208, including, for example, a heat sink 264, one or more fans 268, an autonomous computer 270, and a communication module 272. The electronics plate 260 can include other electrical or mechanical components as well.

[0103] When the camera module 208 is coupled to the housing 204, it will be appreciated that the baseline distance between pairs of the dual EO image sensors and the EO / IR cameras 216, 220 is well-toleranced and is maximized as much as possible. In turn, the delta is consistent and the vision system 200 enables downstream ranging. When in use, and the vision system 200 is operational, the vision system 200 can, for example, have any of the fields of vision described in greater detail in U.S. Provisional Application No. 63 / 742,533, titled “Perception Hardware Configurations for Maritime Vehicle” and filed January 7, 2025, the contents of which are hereby incorporated by reference herein.Example Computing Environment and System

[0104] FIG. 3A depicts an example computing environment 300 in which various techniques and / or examples of the present disclosure can be implemented. The computing environment 300 may perform at least some of the method 100 in accordance with various examples described herein. In some examples, the computing environment 300 (including server 310, computing device 320, optical device 330, and network 340), in addition to the perforated calibration board 112, background 122, attachment 124, and / or other components, may comprise a computing system. It should be understood that the computing system may include additional, fewer, and / or alternative components than that which is described herein. For example, the computing environment 300 / computing system may be a single component / device (e.g., computing device 320). In other examples, the components of the vision system 200 may be or comprise a part of the computing environment 300 / computing system (e.g., utilize autonomous computer 270 to perform the method 100). Advantageously the environment 300 enables greater calibration accuracy and speed with reduced calibration time, improved image quality after distortion correction, and overall increased reliability of vision system calibration.PATENT33894 / H011

[0105] The computing environment 300 may include server 310, computing device 320, optical device 330, and network 340. It should be appreciated that, while the server 310, computing device 320, and optical device 330 are illustrated in FIG. 3A as single components, the example computing environment 300 may include multiple (e.g., dozens, hundreds, thousands) of servers 310, computing devices 320, and / or optical devices 330 (e.g., an EO camera and an LWIR camera). Further, it should be understood that while components and modules are illustrated as distributed across the computing environment 300 that a single component (e.g., server 310, computing device 320, autonomous computer 270) may perform the functionality described herein.

[0106] The server 310 may be associated with an organization (e.g., a company) and / or entity (e.g., the government, military, police, etc.) that calibrates vision systems. The server 310 may include processor 312, memory 314, geometry detection module 314a, parameter estimation module 314b, and networking interface 316. The optical device 330 may be associated with an organization and / or entity that calibrates vision systems. The optical device 330 may be a pail of vision system 200, for example. The optical device 330 may be an electro-optical (EO) camera, infrared (LWIR) camera, etc. The computing device 320 may be associated with an organization and / or entity that calibrates vision systems. The computing device 320 may include processor 322, memory 324, user interface module 324a, networking interface 326, and EO device 328. The network 340 is generally configured to facilitate communication among and / or between the components of the computing environment 300 and / or other components (e.g., via the internet).

[0107] In the example of FIG. 3 A, the server 310 performs at least some of the functionalities and techniques disclosed herein, such as calibrating the optical device 330. The server 310 may include only one server, or multiple servers that are co-located and / or remotely distributed. The server 310 may be part of a cloud network or may otherwise communicate with other hardware or software components within one or more cloud computing environments to send, retrieve, or otherwise analyze data and / or information described herein. In some examples, the computing environment 300 comprises an on-premises computing environment, a multi-cloud computing environment, a public cloud computing environment, a private cloud computing environment, and / or a hybrid cloud computing environment. Advantageously the present techniques may utilize off- site or cloud-based processing thereby improving the scalability and flexibility of deploying such calibration systems especially in the field. The server 310 includes processor 312, memory 314, and / or networking interface 316.PATENT33894 / H011

[0108] The processor 312 includes any suitable number of processors and / or processor types. In some examples, the processor 312 includes one or more central processing units (CPUs), one or more graphics processing units (GPUs), one or more tensor processing units (TPUs), one or more field-programmable gate arrays (FPGAs), one or more application- specific integrated circuits (ASICs), and / or the like. Generally, the processor 312 comprises hardware configured to execute instructions (e.g., processor-executable code / instructions) stored in the memory 314. Advantageously, by utilizing parallel processing architectures (e.g., GPUs and / or TPUs in the computing environment 300) one or more calibration algorithms may execute concurrently on large image datasets. This hardware acceleration not only diminishes processing latency but also ensures that iterative steps (e.g., adaptive thresholding and sub-pixel refinement) are performed with greater speed, enabling real-time correction even under dynamic imaging conditions.

[0109] The memory 314 includes any suitable memory type(s), including one or more volatile memories (e.g., dynamic and / or static random-access memory (RAM)) and / or non-volatile memories (e.g., read-only memory (ROM), erasable programmable ROM (EPROM), electrically EROM (EEROM), NAND flash, and / or solid state drive(s) (SSD(s))), all or any of which are examples of non-transitory computer-readable media. In some examples, the memory 314 stores one or more of: an operating system; one or more software components (e.g., firmware, application(s), binary, source code, executable instructions, machine-learned model(s)); transient data and / or code loaded and / or operated on by one or more software component(s); and / or other suitable components / data. In some examples, the memory 314 stores the geometry detection module 314a and the parameter estimation module 314b. The memory 314 may additionally or alternatively be and / or store one or more databases (e.g., for storing images).

[0110] The geometry detection module 314a generally processes and analyzes images. The geometry detection module 314a in conjunction with the parameter estimation module 314b and image correction module 324b may provide an image correction solution / technique enabling a user to calibrate the optical device 330. The geometry detection module 314a, when executed by the processor 312, facilitates one or more multi- spectral calibration functions, such as processing the plurality of images to analyze the perforated calibration board 112, for example.Advantageously the present techniques provider higher accuracy and shorter computational times using adaptive thresholding, noise reduction, and iterative sub-pixel refinement to improve calibration precision. Further, efficient calibration may be achieved in the various modes ofPATENT33894 / H011 operation (manual, auto comers, contour mode) improving detection accuracy and reducing computational load.

[0111] The geometry detection module 314a may include the Python script 350 illustrated in FIG. 3B. In the example of FIG. 3B, the script 350 may take as input a variable specifying one or more images to be analyzed. The script may process the one or more images and convert the one or more images into one or more grayscale images. The script 350 may run in one of three modes for geometry detection: (1) manual comers, (2) auto corners, or (3) contour. The script 350 may overlay the geometry detection on the one or more original images and cause the one or more overlayed images to be displayed for a user to review. Advantageously, the computing environment 300 leveraging hardware accelerators (e.g., GPUs and / or field-programmable gate arrays) performs complex iterative algorithms (e.g., sub-pixel comer refinement) with low latency contributing to the overall reduction in computational time, ensuring that calibration corrections are updated in real time even for data-intensive applications.

[0112] In one example, the optical device 330 may capture one or more images of a calibration object (e.g., perforated calibration board 112, an IR fiducial) at one or more distances and / or angles. The optical device 330 may transmit the images via network 340 to the server 310. The server 310 may receive a user input (e.g., from user interface module 324a) selecting (1) manual comers mode, for example. The server 310, using the geometry detection module 314a, may cause processor 312 to execute script 350. The script 350 may convert the input images (e.g., captured by optical device 330) to grayscale. The server 310 may cause the script 350 to use the Harris corner detection method to find corner points. The Harris comer detection algorithm may identify regions in the images with large variations in intensity in all directions (e.g., characteristic of comers). The script 350 may dilate the resulting images by increasing the size of the white regions (or foreground) in the images. The script 350 may dilate the resulting images to accentuate the features detected by the Harris comer detector (e.g., making the corner regions more pronounced). The script 350 may threshold the dilated images to convert the images into binary images where only the most prominent features (e.g., corners) are retained. The threshold may be set to 1% of the maximum value in the images, such that only the regions with the highest comer response values (e.g., indicating strong corners) are retained. The thresholding may result in binary images where the pixels corresponding to the detected corners are set to 255 (white), and all other pixels are set to 0 (black). The script 350 may find centroids using the thrcsholdcd images in order to find all connected components (blobs) and theirPATENT33894 / H011 centroids. Because the thresholded images contains isolated white regions corresponding to comers, the script 350 finds the center points of these corner regions. The script 350 uses the centroids obtained as initial estimates for the actual comer locations. However, these estimates may not be precise enough for applications requiring high accuracy. Therefore, the script 350 may refine these comer locations to sub-pixel accuracy. The script 350 iteratively improves the comer estimates based on the local image gradients and a specified termination criteria (e.g., either 100 iterations or a movement of less than 0.001 pixels). The script 350 may overlay the original and refined corner locations onto the original and / or grayscale images for visualization. The script 350 may cause the overlayed images to be displayed for a user (e.g., via the I / O device 328). In the example of FIG. 3C the blue dots 360 are the initial comer estimates and the green dots 370 are sub-pixel refinements.

[0113] In another example, the server 310 may receive a user input (e.g., from user interface module 324a) selecting (2) auto comers mode. In (2) auto corners mode, the script 350 generally attempts to find corners using a chessboard pattern recognition method. The script 350 employs various image processing techniques such as adaptive thresholding and normalization to improve the detection of chessboard comers under different lighting conditions and at different scales. Substantially as before, the optical device 330 may capture one or more images of the calibration object at one or more distances and / or angles and transmit the images via network 340 to server 310. The server 310 using the geometry detection module 314a may cause processor 312 to execute script 350. The script 350 may convert the input images (e.g., captured by optical device 330) to grayscale. The script 350 may take as input (e.g., via user interface module 324a) the size of the chessboard pattern to be detected, specified as the number of inner comers per chessboard row and column. The script 350 may use a combination of flags to control the behavior of the comer detection algorithm. For example, the script 350 may use adaptive thresholding to convert the images to binary (e.g., for handling varying lighting conditions across the image). The script 350 may normalize the images brightness and contrast (e.g., to improve the detection accuracy). The script 350 may run an exhaustive search to find the best pattern detection (e.g., to increase the detection accuracy). The script 350 may filter out quadrangles that do not resemble a chessboard comer. If the chessboard corners are successfully detected the script 350 may overlay the detected corners on the original and / or grayscale images. The script 350 may cause the overlayed images to be displayed to a user. FIG. 3D illustrates an example image of a checkerboard pattern 380 for use with the (2) auto corners mode of the script 350.PATENT33894 / H011

[0114] In yet another example, the server 310 may receive a user input selecting (3) contour mode. In (3) contour mode, the script 350 may generally detect contours in the input images. The script 350 may recognize the curves joining all the continuous points along a boundary which have the same color or intensity. As before, the optical device 330 may capture one or more images of a calibration object at one or more distances and / or angles and transmit the images via network 340 to server 310. The server 310 using the geometry detection module 314a may cause processor 312 to execute script 350. The script 350 may convert the input image to grayscale, so the grayscale image is used as a base for applying a threshold or canny edge detection before finding contours. The script 350 may retrieve all contours without establishing any hierarchical relationships. The script 350 may organize contours into a parent-child hierarchy. The script 350 approximates contours by compressing horizontal, vertical, and diagonal segments of the contour into their end points only (e.g., reducing the amount of information needed to describe the contour). After detecting the contours, the script 350 may overlay the contours on the original and / or grayscale images. The script 350 may cause the overlayed image to be displayed to a user.

[0115] The parameter estimation module 314b generally estimates parameters and establishes a correction in accordance with the calibration method 100. The parameter estimation module 314b in conjunction with the geometry detection module 314a, and image correction module 324b may provide an image correction solution / technique enabling a user to calibrate optical device 330. The parameter estimation module 314b, when executed by the processor 312, facilitates one or more calibration functions, such as establishing a correction by estimating parameters of the optical device 330 based on the analysis of the perforated calibration board 112.

[0116] The parameter estimation module 314b may include a calibration program (e.g., an API, library, algorithm, package, etc.) such as OpenCV, MATLAB Camera Calibration Toolbox, Robot Operating System Calibration Tools, Bundler, COLMAP, Metrical, etc. The calibration program may be a proprietary solution (i.e., developed for a particular use case). In general, the calibration program may establish correspondences and estimate parameters of optical device 330. In general, the calibration program may establish correspondence by matching features across multiple images or with known 3D points (e.g., between the detected 2D image comers with their corresponding 3D world coordinates). The calibration program may estimate parameters of the optical device 330 by applying a mathematical algorithm to estimate intrinsicPATENT33894 / H011 parameters (focal length, principal point, distortion coefficients) and / or extrinsic parameters (position and orientation of the camera in 3D space).

[0117] In one example, the calibration program may establish a correction by estimating parameters of the optical device 330 based on one or more of the geometry detections of geometry detection module 314a. For example, the calibration program may receive as input the geometry detection of script 350. The calibration program may utilize two sets of points: object points (the actual, known coordinates of the calibration object geometry in 3D space) and image points (the coordinates of the detected geometry within the 2D images). The object points are usually predefined based on the known dimensions and geometry of the calibration object, while the image points are obtained from the geometry detection process. With the object points and image points from multiple images, the calibration program may compute the camera's intrinsic and extrinsic parameters. The intrinsic parameters may include the focal length, the principal point, and the distortion coefficients. The extrinsic parameters may describe the camera's position and orientation in a particular environment (e.g., the world). The calibration program may establish a mathematical model to describe how the camera captures a 3D scene and projects it onto a 2D image. For each pixel in the distorted image, the calibration program may use the pixel coordinates (e.g., u, v) and converts the pixel coordinates to normalized coordinates (x, y). This normalization accounts for the optical device’s principal point and focal length. The intrinsic and / or extrinsic parameters of the optical device 330 allow the calibration program to establish a correction (e.g., the radial distortion coefficients kl, k2, k3, etc. and tangential distortion coefficients pl, p2, etc.) for subsequent images captured by optical device 330. The parameter estimation module 314b may transmit one or more established corrections (e.g., to the user interface module 324a, optical device 330, etc.).

[0118] The parameter estimation module 314b may also enable rectification. In one example, using two images captured by a stereo camera (e.g., two IR cameras with overlapping fields of view) the parameter estimation module 314b may use corresponding points to analyze the epipolar geometry (e.g., between the two views of the IR cameras). The parameter estimation module 314b may calculate a fundamental matrix encoding the relationship between the epipolar lines in the images. This may help determine the relative orientation and position of the stereo cameras. The parameter estimation module 314b may estimate the rectification homography to find a transformation (e.g., a homography matrix) that maps each image onto a common plane where the epipolar lines become parallel and horizontal. The image correction module 324b mayPATENT33894 / H011 analyze the relative pose (e.g., position and orientation) of the stereo cameras, which may be determined through camera calibration and / or by analyzing corresponding points in the images. The homography may be calculated to ensure that the corresponding epipolar lines in the transformed images become collinear and parallel to the horizontal axis of the image plane. The parameter estimation module 314b may transmit the homography to the image correction module 324b to correct subsequent images.

[0119] It is to be understood that the rectification techniques described herein may, in some examples, occur after undistortian (e.g., removing lens distortion) as described below in reference to the image correction module 324b. Further it should be understood that while the functionality / components of FIG. 3A may be portrayed as separate components (e.g., the server 310, optical device 330, and computing device 320), the computing environment 300 may in some examples incorporate the components and / or functionality into a singular computing device and / or system (e.g., the computing device 320 detects geometries, estimates parameters, and applies corrections to subsequent images).

[0120] The networking interface 316 comprises one or more hardware components to generally enable the server 310 to communicate via one or more network(s) (e.g., network 340) with other components and / or devices of the computing environment 300, such as the computing device 320, the optical device 330, the server 310 itself (e.g., between components of a server, between two or more servers of two or more servers composing the server 310), and / or other suitable devices or combinations thereof. More specifically, the networking interface 316 enables the server 310 to communicate with any component of the example computing environment 300 across the network 340. The networking interface 316 may comprise hardware and / or software that operates according to at least one communication protocol of the network 340.

[0121] The optical device 330 performs at least some of the functionalities and techniques disclosed herein, such as capturing a plurality of images of the perforated calibration board 112 at a plurality of positions and capturing subsequent images, for example. The optical device 330 may have associated intrinsic parameters (e.g., focal length, principal point, distortion coefficients, etc.) and extrinsic parameters (position, rotation, translation, etc.). The optical device 330 may be or include one or more of an electro-optical (EO) camera, near-infrared (NIR) camera, short-wave infrared (SWIR) camera, mid-wave infrared (MWIR) camera, long-wave infrared (LWIR) camera. In one example, the optical device 330 may be or include a stereo camera arrangement (e.g., two EO cameras). The stereo camera arrangement may include aPATENT33894 / H011 variety of suitable sensors that are capable of capturing light including UV, visible light, and infrared (IR) wavelengths. In another example, the optical device 330 may be an NIR camera. The NIR camera may include a variety of suitable sensors that capture light beyond the visible spectrum, such as between 0.7 and 1.4 micrometers for example. In another example, the optical device 230 may be a SWIR camera. The SWIR camera may include a variety of suitable sensors that capture SWIR wavelengths, from 0.7 to 2.5 micrometers for example. In yet another example, the optical device 330 may be an MWIR camera. The MWIR camera may include a variety of suitable sensors that capture MWIR wavelengths, from 3 to 5 micrometers for example. In a further example, the optical device 330 may be an LWIR camera. The LWIR camera may include a variety of suitable sensors that capture LWIR wavelengths, from 8 to 14 micrometers for example. It is to be understood that the wavelength ranges described herein are non-limiting and, in some examples, may be different and / or overlap.

[0122] In some examples, the computing device 320 may include a computer (e.g., desktop computer, laptop computer, terminal), a mobile device (e.g., smart phone, tablet, a wearable smart device, etc.), an loT device (e.g., a smart speaker, a smart security device, etc.) and / or other suitable computing device. In some examples the computing device 320 may include aspects of one or more of the foregoing in combination, such as a smartphone tablet combination (e.g., a phablet) or a tablet laptop combination (e.g., a 2-in-l tablet laptop), etc. The computing device 320 includes a processor 322 (e.g., similar to the processor 312) and a memory 324 (e.g., similar to the memory 314) for storing and executing one or more software components, computer-executable instructions, etc. The computing device 320 may further include a networking interface 326 (e.g., which may be the same as or similar to the networking interface 338) and an I / O device 328 (e.g., a display, such as a monitor; a user input device, such as a keyboard, mouse, trackpad, gesture and / or biometric tracking device, or the like). The computing device 320 may access services, devices, and / or components of the computing environment 300 via the network 340. In some examples, the computing device 320 transmits and / or receives information and / or data from the server 310 and / or the optical device 330 associated with the vision system calibration techniques described herein. It should be appreciated that, while computing device 320, is illustrated in FIG. 3A as a single component, the computing device 320 may include multiple (e.g., dozens, hundreds, thousands) computing devices 320.PATENT33894 / H011

[0123] The user interface module 324a generally enables users to interact with the dcviccs / componcnts of the environment 300. The user interface module 324a may provide a centralized platform with a user interface (UI) enabling a user to calibrate the optical device 330. The user interface module 324a may be an application developed and / or managed by a third- party. The user interface module 324a, when executed by the processor 322, facilitates one or more vision system calibration functions, such as positioning a perforated calibration board at a plurality of positions relative to the optical device (e.g., by causing the optical device 330 and or the perforated calibration board 112 to move), and capturing, via the optical device 330, a plurality of images of the perforated calibration board 112 at a plurality of positions, for example. In some examples, the user interface module 324a may transmit or cause to be transmitted one or more images from the optical device 330 to the server 310. In other examples, the user interface module 324a may cause the image correction module 324b to apply an established correction to subsequent images captured by the optical device.

[0124] The image correction module 324b generally applies image corrections (e.g., a distortion model) to subsequent images taken by the optical device 330. The image correction module 324b may apply the established correction (e.g., the distortion coefficients) to subsequent images, removing distortions such as radial and tangential lens distortions. The image correction module 324b may use the distortion coefficients (e.g., kl, k2, pl, p2, etc.) to calculate how much a point is distorted. The image correction module 324b may undistort the image by applying the inverse of the distortion model to get the undistorted normalized coordinates (e.g., x', y'). The image correction module 324b may convert the undistorted normalized coordinates (e.g., x', y') back to pixel coordinates (e.g., u', v') in the corrected image. This results in images that more accurately represent the observed scene, improving the accuracy of optical device 330. The image correction module 324b may include verification of the correction by capturing new images of the calibration object with known dimensions and comparing the observed measurements with the expected values (e.g., a threshold for a calibration accuracy metric such as reprojection error, mean reprojection error (MRE), root mean square (RMS) reprojection error, epipolar error, standard error of intrinsic parameters, standard error of extrinsic parameters, inter-point distance error, alignment error, calibration pattern residual.). The image correction module 324b may allow for adjustments to the correction if discrepancies are found (e.g., iteratively refine the calibration in real time until a threshold calibration accuracy metric is reached).PATENT33894 / H011

[0125] In some examples, the image correction module 324b may enable rectification by warping images. The image correction module 324b may receive from the parameter estimation module 314b a rectification homography. The image correction module 324b may warp (e.g., transform) the images using the rectification homography. The image correction module 324b may reproject the images onto the common plane, effectively aligning them as if they were captured by cameras with parallel optical axes. After rectification, the two images may appear as if they were taken by cameras with parallel optical axes, and the epipolar lines become horizontal such that corresponding points now lie on the same horizontal line in both images, reducing the search space from 2D to ID.

[0126] It should be understood that, in some examples, each image may be undistorted to remove lens distortion prior to the parameter estimation module 314b determining a rectification homography. This undistorian may ensure that straight lines in the real world appear straight in the images, making subsequent alignment more accurate. Advantageously the integration of the user interface and image correction modules allow users to interact with the system in a responsive and intuitive manner leading to rapid feedback and improved ease-of-use in industrial or field calibration scenarios. Further, the described techniques may utilize modular software design to interface with various hardware components (e.g., using advanced networking protocols) to provide an improved calibration workflow, reducing manual intervention and errors.

[0127] The network 340 includes wired and / or wireless communication network(s) such as a cellular network (e.g., 5G®, 4G LTE®, 3G®), a Wi-Fi® network (802.11 standards), a microwave access network (e.g., WiMAX®), and / or any other suitable wide area network (WAN), local area network (LAN), personal area network (PAN), etc. Moreover, the network 340 may be a single communication network, or may include multiple communication networks of one or more types (e.g., one or more wired and / or PANs or LANs, and / or one or more WANs such as the Internet). In some examples, the network 340 includes multiple, entirely distinct networks (e.g., one or more networks for communications between server 310 and computing device 320, and a separate, Bluetooth® or wireless LAN (WLAN) network for communications between server 310 and computing device 320, and so on). It should be appreciated that, while the network 340 is illustrated in FIG. 3A as a single component, the network 340 may include multiple (e.g., dozens, hundreds, thousands) networks 340. Advantageously, the integration of dedicated high-speed networking interfaces and enables seamless data transfer from variedPATENT33894 / H011 spectral channels enabling dynamic cross-verification of calibration parameters while also providing resilience against rapidly changing environmental variations (e.g., outdoors) with rapid relay and processing of sensor data in real time.

[0128] It will be appreciated that each of the one or more components may be implemented as hardware (e.g., processor, die, integrated device), software (e.g., non-transitory processor readable medium), and / or combinations thereof, in one or more devices (e.g., processor, chip, computer, tablet, mobile device).IR Fiducial (First Example)

[0129] FIGS. 4A and 4B illustrate another example of a calibration object 400 that can be used instead of or can incorporate the perforated calibration board 112. In this example, the calibration object 400 takes the form of an IR fiducial that includes a pair of boards (which may also be referred to as a pair of plates, sheets, panels, etc.) including a forward board 410 and a rear board 420 coupled to the forward board 410. In some examples, the forward board 410 can be identical or similar to the perforated calibration board 112, in which case the forward board 410 can include square-shaped perforations 416 like the perforations 116. The rear board 420 is generally different from the forward board 410. In some examples, the rear board 420 may be a calibration board that has different or no perforations and defines a background (e.g., the background 122). Each of the boards 410, 420 has a front side 412, a back side 414, and a thickness t (e.g., ~2 mm) defined between the front side and the back side. The forward and rear boards 410, 420 are coupled to one another and separated by one or more spacers 430 secured to the boards 410, 420 (e.g., in the comers) with fasteners (e.g., bolts, screws, studs, etc.). In this example, each spacer 430 is a spacing post 430. The spacing posts 430 define a distance d (e.g., ~25 mm) between the boards 410, 420 sufficient to ensure thermal isolation between the forward board 410 and the rear board 420. It will be appreciated that, like the perforated calibration board 112, the boards 410, 420 may be made of any suitable material such as but not limited to aluminum, PVC, soda lime glass, or another suitable material. Similarly, the calibration object 400 is not limited to the square shaped perforations 416 depicted in FIGS. 4A and 4B, but like the calibration board 112 may include any suitable arrangement and / or geometries of shapes, sizes, etc.

[0130] In some examples, the rear board 420 may be heated by coupling a heating element (not shown) to the rear board 420 (e.g., to the back side of the rear board 420). The heating element can, for example, take the form of pipe heating tape, a heating cable, a heat trace cable, a heaterPATENT33894 / H011 wire / cord / rope, resistance wire, a self-regulating heating cable, a positive temperature coefficient heater, etc. Thus, when the rear board 420 is heated, the forward board 410 may be cold relative to the real- board 420. As partially depicted in FIG. 4A, the pair of boards 410, 420 may be spray-painted (e.g., black) on their front sides and left plain (or unpainted) on their back sides. The paint on the front sides of each of the boards 410, 420 beneficially breaks reflections that may normally cause the aluminum to have a highly variable signature in LWIR. Because the forward board 410 is not heated, the bare material (e.g., aluminum) on the rear of the forward board 410 is able to dissipate any incident heat rapidly, maintaining the forward board 410 at ambient (e.g., room / environment) temperature. The paint on the front of the rear board 420 acts as a thermal bank and diffuser, ensuring a uniform signature in LWIR and permitting the rear board 420 to glow long after the heating element coupled the back of the rear board 420 has been turned off and the back of the rear board 420 has returned to nearly ambient temperature.

[0131] Turning on the heating element for less than one minute (e.g., making the rear board 420 warm to the touch) is sufficient to make the rear board 420 glow for over 10 minutes in IR. As depicted in FIG. 4C, the rear board 420 may glow through the perforations of the forward board 410 shortly after turning off the heating element. As illustrated in FIG. 4D, the rear board 420 still retains heat approximately 5 minutes after turning off the heating element. As illustrated in FIGS. 4E and 4F, the features of the calibration object 400 remain clearly distinguishable with very high contrast even in proximity to other bright objects, such as people.

[0132] As illustrated in FIGS. 4G-4R, the forward board 410 may instead have circular perforations 418 of different sizes. In some examples, the forward board 410 may be painted black, and the rear board 420 may be painted white. Even in a poorly lit environment the contrast on the circles may be sufficient to detect circles (e.g., using the calibration method 100). In some examples, white paint may not have many significant differences from black in LWIR. The reflectance and emissivity of the back side of the forward board 410 may ensure that the forward board 410 stays relatively cool relative to the rear board 420. In one example, the color scheme may be reversed. Beneficially this may remove contrast reduction due to shadows on the rear board 420.

[0133] In one example, the spacing posts 430 may be approximately 1 inch long, such that the forward and rear boards 410, 420 are approximately 1 inch apart from one another. In another example, e.g., in the images of FIGS. 4G, 4J, 4P, & 4Q, the spacing posts 430 may instead be —1 / 2 inch. This example may not change many of the basic LWIR properties, so this beneficiallyPATENT33894 / H011 increases robustness and reduces weight. Advantageously, the edges 419 of the forward board 410 and the respective perforations 418 may not show even at sharp angles as illustrated in the images of FIGS. 4G and 4 J.

[0134] In some examples, the calibration object 400 may include or be coupled to a thermal control system (not shown). The thermal control system may enable a user to manually and / or automatically control (e.g., adjust or shut off) the heating element (e.g., pipe heating tape) as necessary to control the temperature of the rear board 420. In some examples, to enable automation and ensure safety the thermal control system may automatically maintain a temperature difference (e.g., 5°C-10°C ) between the boards 410, 420. The boards 410, 420 may include one or more temperature sensors to measure the relative temperature difference between the boards. The thermal control system may include a microcontroller to run a bang-bang controller or similar controller (e.g., on-off controller, hysteresis controller, relay controller, threshold controller, pulse-width modulation, sliding mode control, etc.) to control the heating element as desired. The thermal control system may include a safety cutoff. In some examples, the desired safety cut off may be ~55°C to avoid injuring a user manually positioning the calibration object 400. The thermal control system may enable the rear board 420 to maintain detectability in IR for a duration exceeding the active heating period.

[0135] The calibration object 400 may also have interior lighting for the pair of boards. Because an EO camera may rely upon ambient light, there may be many shadows between the boards 410, 420, as illustrated in FIG. 4R. These shadows may be especially problematic under harsh bright conditions, when the shadows have better contrast than edges of the calibration object 400. Thus, the pair of boards 410, 420 may include one or more lighting elements (not shown) affixed to the forward board 410 and facing rearward, towards the rear board 420. The interior lighting elements may be thermally isolated from the forward board 410 by ceramic, Kapton, MPET, or one or more other suitable materials to minimize heat transfer. In some examples, the lighting elements may be light-emitting diodes (LEDs) with an appropriate diffuser. In some examples, a matte / flat paint may help with thermal isolation. Additionally or alternatively, the boards 410, 420 may be painted white on the front side and black on the back side such that shadows may not affect contrast as much.

[0136] As discussed, the thermal insulation of portions of the pair of boards 410, 420 may be desirable. Although examples depicted in the various FIGs may use highly thermally conductive spacing posts 430 and / or fasteners between the boards 410, 420, the conditions under which thePATENT33894 / H011 images of the IR fiducials were captured may not cause issues due to the highly emissive nature of the material (e.g., aluminum) and the relatively cold temperatures under which they operated (e.g., indoors). The fastener locations at the comers may be clearly visible as illustrated in FIG. 4S but may not cause contrast issues for the main features (e.g., the perforations 416, 418) of the fiducial. In some examples with less controlled circumstances (e.g., a warmer or colder environment), the issue may be more significant. Thus, the spacing posts 430 may instead be made of ceramic, garolite, polyvinyl chloride (PVC), polyoxymethylene (POM), fiberglass, polyetheretherketone (PEEK), nylon, or a similar insulating material. A similar’ feature may be utilized for the interior lighting described herein in order to direct heat from the lighting elements (e.g., the LEDs) towards the rear board 420. Because aluminum is the substance of choice for many space-blanket materials, contact transfer through conduction may be a concern. As such, an insulator backing (e.g., ceramic, an air gap, etc.) may be used to mitigate heat transfer.

[0137] When the pair of boards 410, 420 is captured past a certain angle as illustrated in FIG. 4T, the rear board 420 (which serves as the heating board) may not be visible through the perforations 418 of the front board 410 (e.g., the perforated target calibration board 112). Thus, to ensure calibration may be done at a variety of angles, the spacer 430 may be a side board connecting the edges of the pair of boards 410, 420 (and thereby connecting the forward board 410 to the rear board 420) and defining a shadowbox (i.e., essentially enclosing the space between the forward board 410 and the rear board 420). In one example, the side board may be made of aluminum allowing it to heat up. The side board may connect the forward board 410 and rear board 420 across an area greater than the spacing posts 430 and as such the forward board 410 may include an additional insulation layer. Adding insulating edges to the forward board 410 may prevent heat leakage. The inside of the defined shadowbox may be coated with metallized polyethylene terephthalate (MPET), Kapton or other similar material, which may effectively reflect the rear board 420, thereby minimizing heat transfer to the forward board 410.

[0138] In some examples, the boards 410, 420 may include a slot (not shown) for a hot plastic board to be inserted. The plastic board may provide substantially similar functionality to the heated rear board 420. The plastic board may be heated separately from the remaining fiducial, and inserted into the slot (e.g., when calibrating an optical device passively in an indoor environment). A polymer coating (e.g., provided by spray paint) may be used on the front side of the rear board 420. Because a reliable heat source (e.g., thermal chamber or low-temperaturePATENT33894 / H011 oven) may be required, this approach may not be suitable for uncontrolled environments (e.g., field calibration).IR Fiducial (Alternative Examples)

[0139] FIGS. 5A-5C generally illustrate a plurality of calibration objects 500 mounted together using a mount 502 (which may be referred to as a stand, platform, base, frame, or rack) and mounting hardware 504 (including but not limited to one or more arms, brackets, supports, fasteners, and fixtures). The mount 502 may be stationary (e.g., placed on and / or coupled to the ground and sufficiently weighed down) and / or moveable (e.g., placed on and / or coupled to a wide flat metal cart with castor wheels). Each of the calibration objects 500 illustrated in FIGS. 5A-5C includes a plurality of surfaces (e.g., multiple board layers like front board 410, rear board 420, etc.), having various thermo-optical characteristics, and, in some examples, including thermal control components, as will be discussed in greater detail below. In some examples, the calibration objects 500 and the LWIR techniques described herein enable field (e.g., off-site, rugged terrain, remote, etc.) calibration. In other examples, the calibration objects 500 and the LWIR techniques described herein may facilitate in-house calibration (e.g., on-site, during maintenance, after manufacture, etc.)

[0140] Beneficially, the mount 502 allows the plurality of calibration objects 500 mounted thereto to be mounted in any number of different positions relative to one another and to the environment surrounding the mount 502. For example, FIG. 5A depicts a static arrangement of calibration objects 500 mounted to the mount 502 but missing two calibration objects 500 in the middle row. The other calibration objects 500 depicted in FIG. 5A have been moved from their nominal positions in which the calibration objects 500 are aligned with one another (e.g., such that the calibration objects 500 in each row and column are level and uniformly spaced). In some examples, such as the example illustrated in FIG. 5A, the mounting hardware 504 can include a plurality of movable arms each mounted to both the mount 502 and one of the calibration objects 500 so as to allow the calibration objects 500 mounted to the mount 502 via the movable arms to be adjusted or moved relative to one another and to the environment in situ. Alternatively, as illustrated in FIGS. 5B and 5C, the calibration objects 500 can be mounted together using the mount 502 and the mounting hardware 504 so that the calibration objects 500 are arranged in a compact 3 x 3 grid, matrix, array, arrangement, or layout. Regardless of the exact arrangement of the calibration objects 500, the plurality of calibration objects 500 may be used to calibrate the extrinsic properties of one or more imaging devices (e.g., LWIR camera)PATENT33894 / H011 with non-overlapping field of views (FOVs) relative to each other. This may facilitate extrinsic cross-calibration between cameras operating on different spectra without requiring overlapping fields of view, and may further support approaches for establishing a unified vision model for the system.

[0141] In some examples, one or more of the calibration objects 500 may take the form of the field service representative (FSR) calibration object 530 that is illustrated in FIGS. 5D-5G and enables lightweight, portable multi- spectral calibration. The calibration object 530 may appear as a printed calibration object / board for EO cameras (e.g., as depicted in FIG. IL) while also being LWIR capable. The FSR calibration object 530 generally includes a multi-layer construction including a white front board 532, a thinner middle board 534 painted black for optimized thermal performance and shadow avoidance, and a rear support board 536. All three layers may be thermally insulated from one another. In examples for field use, two layers may be used, with slight protrusions on the side of the rear board to mount handles and a heating system.

[0142] Like the other calibration objects / boards described herein, the white front board 522 includes a plurality of perforations 538. In this example, the plurality of perforations 538 are circular in shape. In some examples, the white front board 532 may have a thickness of approximately 2 mm to provide rigidity. The middle board 534 may be thermally insulated from the rear board 536 and include a low-profile heating pad (e.g., a heating element 539) that causes the calibration object 500 to appear to glow in LWIR, thereby enabling and improving calibration precision across spectra. In this example, FSR calibration object 530 also includes a pair of handles 540 coupled to opposing sides of the rear support board 536 to allow the FSR calibration object 530 to be moved, rotated, etc. from one location, orientation, position, etc. to another. The handles 540 are preferably thermally insulated. The FSR calibration object 530 of FIGS. 5D-5G therefore offers improved portability and ease of use in field calibration scenarios where manual operation is desired, providing a cost-effective and robust calibration technique.

[0143] It should be understood that the arrangement of large and small circles on the board is provided for descriptive purposes and may include additional and / or alternative arrangements. In some examples, all the mounted calibration objects 500 are uniquely identifiable (e.g., via a QR code, bar code, data matrix code, Aztec code, etc.), to facilitate orientation of multiple cameras relative to multiple targets. Advantageously, the present techniques allow the calibration objects to be oriented in arbitrary directions relative to each other, which improvesPATENT33894 / H011 the calibration process by enabling the viewing of multiple orientations without having to move an optical device and / or calibration object.

[0144] As best illustrated in FIG. 5G, the FSR calibration object 530 includes or is coupled to a thermal control system 550 for controlling the temperature of the various components of the FSR calibration object 530. In this example, the thermal control system 550 is mounted to the middle board and rear boards 534, 536, such that the thermal control system 550 extends outward (rearward) from the rear board 536. The thermal control system 550 generally includes the heating element 539 described above, which is powered by a battery 552 (e.g., a rechargeable 18V unit), a low-power microprocessor 556 connected to the battery 506 through a step-down transformer 554, a relay 559 connected to the low-power microprocessor 556 and the heating element 539, and one or more temperature sensors 558 affixed to the front board 532 and / or the rear board 536 of the calibration object 500 and connected to the low-power microprocessor 556.

[0145] In some examples, the FSR calibration object 530 may include a holding element (e.g., a battery mount, receptacle, etc.) for the battery 552. In some examples, the holding element for the battery 552 is mounted on the rear board 536. The battery mount may include mounting hardware, an on / off switch, a fuse holder, and one or more wire terminals. In other examples, the battery 552 may be a permanently mounted charging solution (e.g., wired). The low-power microprocessor 556 may control operation of the heating element 539 based on desired thermal characteristics. For example, once a desired maximum temperature is reached (e.g., as measured by the temperature sensors 558), the micro-processor 556 may cause the relay 559 to stop current to the heating element 539. Similarly, once a desired minimum temperature is reached, the micro-processor may cause the relay 559 to provide current to the heating element 539.Advantageously, the thermal control system 550 allows the battery 552 to operate at a lower duty cycle extending the useful lifespan of the calibration object 500.

[0146] One or more of the calibration objects 500 may also or alternatively take the form of the FSR calibration object 560 that is illustrated in FIGS. 5H-5K. The FSR calibration object 560 is virtually identical to the FSR calibration object 530 but for the fact that the FSR calibration object 560 has a pair of handles 540 each of which is shorter than the pair of handles 540 of the FSR calibration object 530.Example Target Confirmation

[0147] FIG. 6 generally illustrates a system 600 for confirming that one or more optical devices 602 have been properly calibrated for object detection. The system 600 includes one orPATENT33894 / H011 more target confirmation models 601 , the optical devices 602, a network 604, and a computing device 606. It should be understood that the system 600 may include additional, fewer, and / or alternative components and architectural configurations without deviating from the techniques described herein. Advantageously the system 600, utilizing error checking, statistical correlation, and machine learning based entity recognition confirms the calibration using additional target models and feedback controls, such as computer implemented modules (e.g., calibration confirmation module, user interface module). The system 600 may integrate with the network and processing hardware to automatically verify calibration in real time adding to enhanced, reliable performance across various fields improving field deployments and integration with multi- spectral cameras.

[0148] The target confirmation models 601 are preferably models of real-world or physical target objects (e.g., maritime vehicles, cars, trains, airplanes, people, buoys, environments, etc.) that are used to confirm that the one or more optical devices 602 have been properly calibrated for object detection. In some examples, the target confirmation models 601 may include cutout shapes, manufactured shapes (e.g., made of aluminum or another suitable material), and / or may be constructed with heat-differentiating materials to enable end-to-end testing (and, hopefully confirmation) of proper object detection by the optical devices 602, which may be substantially similar or different than the optical device 330 described above (e.g., LWIR and EO cameras). In some examples, the present system 600 may be used for thermal emissions testing capturing the reflectance and emission of the target confirmation models 601 to create an image (e.g., as depicted on user interface 612a). The target confirmation models 601 may be sufficiently accurate so as to be tested on a variety of optical devices. The target confirmation models 601 may be positioned at a variety of distances from the optical devices 602 substantially similar to or different than the calibration techniques described herein. The target confirmation models 601 may be scaled down models approximately 18 inches wide and positioned a few feet away from the optical devices 602. In some examples, cloth and / or cardboard may be used with the target confirmation models 601 to simulate the surrounding environment (e.g., ocean, grass, sky, etc.).

[0149] The network 604 may be a wired or wireless network of any suitable communication protocol substantially similar to or different than the network 340 described above. The computing device 606 may be a single or plurality of computing devices (e.g., laptop, server, handheld, etc.) substantially similar or different than the server 310 and computing device 320 described above. The computing device 606 may include one or more processors 608 (e.g.,PATENT33894 / H011CPU, GPU, TPU, etc.), networking interfaces 610 (e.g., for Ethernet), memories 612 (e.g., non- transitory CRM), and I / O devices 614 (e.g., monitor, keyboard, mouse, etc.) all substantially similar or different than processor 312, 322, networking interface 316, 326, memory 314, 324, and I / O device 328 described above in reference to FIG. 3A.

[0150] The memory 612 may include the user interface module 612a and a target confirmation module 612b (e.g., a software module stored and / or executed in the system 600). The user interface module 612a may provide a graphical user interface (GUI) for interacting with the target confirmation module 612b. The target confirmation module 612b may extract features (e.g., heat patterns that simulate vehicle engines, human heat signatures, light emissions, etc.) from images of the target confirmation models 601 (e.g., captured by an optical device 602) to ensure comprehensive detection. The target confirmation module 612b may include and / or otherwise access (e.g., via an API) any suitable computer vision, machine learning, and / or image detection techniques, models, algorithms, etc. to facilitate reliable target confirmation of the target models 601.

[0151] It should be understood that, in some examples, the user interface module 612a and target confirmation module 612b may not be stored on the memory 612 and may instead be stored in memory of another computing device (e.g., a remote server, cloud provider, database, etc.).

[0152] In operation, a user (e.g., a field service representative (FSR), technician, quality assurance officer, etc.) may begin a target confirmation by interacting with the user interface module 612a using a mouse, keyboard, and monitor (e.g., RO device 614). The computing device 606 may cause the optical device(s) 602 to capture one or more images of the target confirmation models 601 by transmitting a command via networking interface 610 over the network 604 (e.g., via Wi-Fi®). The optical device(s) 602 may transmit imaging data over the network 604 back to the computing device 606. The target confirmation module 612b receives imaging data from optical device(s) 602 of the target confirmation models 601. The target confirmation module 612b may cause the one or more processors 608 to execute one or more computer vision algorithms (e.g., edge detection, contour mapping, feature extraction, etc.) to isolate and identify key target features (e.g., perforations, beveled edges, geometric markers inherent to the target confirmation models 601, etc.). In some examples, the target confirmation module 612b may cause the one or more processors 608 to execute a machine learning model (e.g., convolutional neural networks) trained on representative datasets (e.g., real-world orPATENT33894 / H011 physical target objects) and analyze these features to determine if they conform with the expected target shape, pattern, size, orientation, heat signature, etc. The target confirmation module 612b may include pattern recognition and statistical correlation to compare detected parameters against predefined target confirmation standards, taking into account both spatial and thermal signatures. Moreover, the system 600 may fuse data from multiple spectra (e.g., visible and LWIR) to improve detection robustness under varying environmental and lighting conditions. The target confirmation module 612b may include a triggering feedback mechanism to confirm the proper alignment and validity of the target models 601 once a confidence score exceeds a predetermined threshold. In some examples, if discrepancies are observed corrective measures may be initiated. The target confirmation module 612b may output a confirmation determination image to user interface module 612a (e.g., depicting proper identification of target confirmation models 601). This system configuration may support target confirmation by optical devices 602 with disjointed fields of view. The present techniques may be particularly advantageous testing across a variety of distances and environments.

[0153] Finally, although certain methods and objects have been incorporated herein by reference, the scope of coverage of this patent is not limited thereto. On the contrary, while the invention has been shown and described in connection with various preferred examples, it is apparent that certain changes and modifications, in addition to those mentioned above, may be made. This patent covers all examples of the teachings of the disclosure that fairly fall within the scope of permissible equivalents. Accordingly, it is the intention to protect all variations and modifications that may occur to one of ordinary skill in the art.Additional Considerations

[0154] Further, although certain multi- spectral calibration and related system, methods, and components have been described herein in accordance with the teachings of the present disclosure, the scope of coverage of this patent is not limited thereto. On the contrary, while the invention has been shown and described in connection with various preferred embodiments, it is apparent that certain changes and modifications, in addition to the mentioned above, may be made. This patent covers all embodiments of the teachings of the disclosure that fairly fall within the scope of permissible equivalents. Accordingly, it is the intention to protect all variations and modifications that may occur to one of ordinary skill in the art.

[0155] Still further, when implemented, any of the methods and techniques described herein or portions thereof may be performed by executing software, one or more non-transitory, tangible,PATENT33894 / H011 computer readable storage media or memories such as magnetics disks, laser disks, optical discs, semiconductor memories, biological memories, other memory devices, or other storage media, in a RAM or ROM of a computer or processor, etc.

[0156] To aid the Patent Office and any readers of any patent issued on this application in interpreting the claims appended hereto, applicants do not intend any of the appended claims or claim elements to invoke 35 U.S.C. 112(f) unless the words “means for” or “step for” are used. Examples

[0157] Moreover, although the foregoing text sets forth a detailed description of numerous different embodiments, it should be understood that the scope of the patent is defined by the words of the claims set forth at the end of this patent. The detailed description is to be construed as exemplary only and does not describe every possible embodiment because describing every possible embodiment would be impractical, if not impossible. Numerous alternative embodiments could be implemented, using either current technology or technology developed after the filing of the patent, which would still fall within the scope of the claims. By way of example, and not limitation, the disclosure herein contemplates at least the following:

[0158] Example 1. A method of calibrating a vision system, comprising: positioning at least one of (i) an optical device or (ii) a perforated calibration board at a plurality of positions relative to the other of (i) the optical device or (ii) the perforated calibration board; capturing, via the optical device, a plurality of images of the perforated calibration board at the plurality of positions; processing the plurality of images to analyze the perforated calibration board; establishing a correction by estimating parameters of the optical device based on the analyzed perforated calibration board; and applying the correction to subsequent images captured by the optical device.

[0159] Example 2. The method of example 1, further comprising: heating a second calibration board to one or more temperatures within an operating range of the optical device; and positioning the second calibration board adjacent the perforated calibration board, wherein the capturing comprises capturing, via the optical device, the plurality of images for each temperature of the one or more temperatures.

[0160] Example 3. The method of examples 1 or 2, wherein the perforated calibration board has (i) at least one bevel defining an edge of the perforated calibration board, (ii) at least one bevel defining one or more of the perforations, or (iii) one or more perforations representative of at least one geometric or non-gcomctric shape.PATENT33894 / H011

[0161] Example 4. The method of any one of examples 1 -3, wherein the optical device detects a color difference between the perforated calibration board of a first color and at least one bevel of a second color that is different from the first color, and wherein the optical device detects the second color as darker than the first color.

[0162] Example 5. The method of any one of examples 1-4, wherein the optical device detects at least one of (i) one or more edges defining a perforation of the perforated calibration board or (ii) one or more edges of the perforated calibration board as substantially a same color as a background.

[0163] Example 6. The method of any one of examples 1-5, wherein the optical device comprises a plurality of cameras, each camera having a sensitivity to one or more spectra, and wherein the cameras arc arranged to provide an overlapping field of view.

[0164] Example 7. The method of any one of examples 1-6, wherein the optical device includes at least one of (a) a near-infrared (NIR) camera, (b) a short-wave infrared (SWIR) camera, (c) a mid-wave infrared (MWIR) camera, (d) a long-wave infrared (LWIR) camera, or (e) an electro-optical (EO) camera, and wherein processing the plurality of images to analyze the perforated calibration board includes associating one or more edges of at least one of (i) the one or more perforations of the perforated calibration board or (ii) the perforated calibration board with a background.

[0165] Example 8. The method of any one of examples 1-7, wherein establishing the correction by estimating parameters of the optical device includes (i) detecting a predetermined pattern formed by perforations of the calibration board, (ii) comparing positions of the detected perforations with a corresponding desired pattern, and (iii) correcting parameters of the optical device based on deviations between the detected pattern and desired pattern.

[0166] Example 9. The method of any one of examples 1-8, wherein processing the plurality of images to analyze the perforated calibration board includes (i) applying an adaptive thresholding algorithm the captured images to generate a binary image accentuating high- contrast calibration features, and (ii) refining detected feature locations to sub-pixel accuracy using an iterative gradient-based refinement algorithm.

[0167] Example 10. The method of any one of examples 1-9, further comprising: capturing, via the optical device, a verification image of the perforated calibration board after applying the correction; analyzing the verification image to determine a calibration accuracy metric; andPATENT33894 / H011 adjusting, in real time, the estimated correction parameters if the calibration accuracy metric docs not meet a predetermined threshold.

[0168] Example 11. A system for calibrating a vision system, comprising: a perforated calibration board; an optical device of the vision system; one or more processors; and one or more memories storing computer-executable instructions that cause the one or more processors to at least: position at least one of (i) the optical devices or (ii) the perforated calibration board at a plurality of positions relative to the other of (i) the optical device or (ii) the perforated calibration board; capture, via the optical device, a plurality of images; process the images to analyze the perforated calibration board; establish a correction by estimating parameters of the optical device based on the analysis of the perforated calibration board; and apply the correction to subsequent images captured by the optical device.

[0169] Example 12. The system of example 11, wherein the computer-executable instructions, when executed by the one or more processors, further cause the system to: heat a second calibration board to one or more temperatures within an operating range of the optical device; and position the second calibration board adjacent the perforated calibration board, wherein the capturing comprises capturing, via the optical device, the plurality of images for each temperature of the one or more temperatures.

[0170] Example 13. The system of examples 11 or 12, wherein the perforated calibration board has (i) at least one bevel defining an edge of the perforated calibration board, (ii) at least one bevel defining one or more of the perforations, or (iii) one or more perforations representative of at least one geometric or non-geometric shape.

[0171] Example 14. The system of any one of examples 11-13, wherein the optical device detects a color difference between the perforated calibration board of a first color and at least one bevel of a second color that is different from the first color, and wherein the optical device detects the second color as darker than the first color.

[0172] Example 15. The system of any one of examples 11-14, wherein the optical device detects at least one of (i) one or more edges defining a perforation of the perforated calibration board or (ii) one or more edges of the perforated calibration board as substantially a same color as a background.

[0173] Example 16. The system of any one of examples 11-15, wherein the optical device comprises a plurality of cameras, each camera having a sensitivity to one or more spectra, and wherein the cameras arc arranged to provide an overlapping field of view.PATENT33894 / H011

[0174] Example 17. The system of any one of examples 1 1-16, wherein the optical device includes at least one of (a) a near-infrared (NIR) camera, (b) a short-wave infrared (SWIR) camera, (c) a mid-wave infrared (MWIR) camera, (d) a long-wave infrared (LWIR) camera, or (e) an electro-optical (EO) camera, and wherein process the plurality of images to analyze the perforated calibration board includes associating one or more edges of at least one of (i) the one or more perforations of the perforated calibration board or (ii) the perforated calibration board with a background.

[0175] Example 18. The system of any one of examples 11-17, wherein establishing the correction by estimating parameters of the optical device includes (i) detecting a predetermined pattern formed by perforations of the calibration board, (ii) comparing positions of the detected perforations with a corresponding desired pattern, and (iii) correcting parameters of the optical device based on deviations between the detected pattern and desired pattern.

[0176] Example 19. The system of any one of examples 11-18, wherein process the plurality of images to analyze the perforated calibration board includes (i) applying an adaptive thresholding algorithm the captured images to generate a binary image accentuating high- contrast calibration features, and (ii) refining detected feature locations to sub-pixel accuracy using an iterative gradient-based refinement algorithm.

[0177] Example 20. The system of any one of examples 11-19, wherein the computerexecutable instructions, when executed by the one or more processors, further cause the system to: capture, via the optical device, a verification image of the perforated calibration board after applying the correction; analyze the verification image to determine a calibration accuracy metric; and adjust, in real time, the estimated correction parameters if the calibration accuracy metric does not meet a predetermined threshold.

[0178] Example 21. One or more non-transitory computer-readable medium having computerexecutable instructions stored thereon that, when executed by one or more processors, cause the one or more processors to at least: position at least one of (i) an optical device or (ii) a perforated calibration board at a plurality of positions relative to the other of (i) the optical device or (ii) the perforated calibration board; capture, via the optical device, a plurality of images; process the images to analyze the perforated calibration board; establish a correction by estimating parameters of the optical device based on the analyzed perforated calibration board; and apply the correction to subsequent images captured by the optical device.PATENT33894 / H011

[0179] Example 22. The one or more non-transitory computer-readable medium of example 21, wherein the computer-executable instructions further cause the one or more processors to: heat a second calibration board to one or more temperatures within an operating range of the optical device; and position the second calibration board adjacent the perforated calibration board, wherein the capturing comprises capturing, via the optical device, the plurality of images for each temperature of the one or more temperatures.

[0180] Example 23. The one or more non-transitory computer-readable medium of examples 21 or 22, wherein the perforated calibration board has (i) at least one bevel defining an edge of the perforated calibration board, (ii) at least one bevel defining one or more of the perforations, or (iii) one or more perforations representative of at least one geometric or non-geometric shape.

[0181] Example 24. The one or more non-transitory computer-readable medium of any one of examples 21-23, wherein the optical device detects a color difference between the perforated calibration board of a first color and at least one bevel of a second color that is different from the first color, and wherein the optical device detects the second color as darker than the first color.

[0182] Example 25. The one or more non-transitory computer-readable medium of any one of examples 21-24, wherein the optical device detects at least one of (i) one or more edges defining a perforation of the perforated calibration board or (ii) one or more edges of the perforated calibration board as substantially a same color as a background.

[0183] Example 26. The one or more non-transitory computer-readable medium of any one of examples 21-25, wherein the optical device comprises a plurality of cameras, each camera having a sensitivity to one or more spectra, and wherein the cameras are arranged to provide an overlapping field of view.

[0184] Example 27. The one or more non-transitory computer-readable medium of any one of examples 21-26, wherein the optical device includes at least one of (a) a near- infrared (NIR) camera, (b) a short-wave infrared (SWIR) camera, (c) a mid-wave infrared (MWIR) camera, (d) a long-wave infrared (LWIR) camera, or (e) an electro-optical (EO) camera, and wherein process the plurality of images to analyze the perforated calibration board includes associating one or more edges of at least one of (i) the one or more perforations of the perforated calibration board or (ii) the perforated calibration board with a background.

[0185] Example 28. The one or more non-transitory computer-readable medium of any one of examples 21-27, wherein establishing the correction by estimating parameters of the optical device includes (i) detecting a predetermined pattern formed by perforations of the calibrationPATENT33894 / H011 board, (ii) comparing positions of the detected perforations with a corresponding desired pattern, and (iii) correcting parameters of the optical device based on deviations between the detected pattern and desired pattern.

[0186] Example 29. The one or more non-transitory computer-readable medium of any one of examples 21-28, wherein process the plurality of images to analyze the perforated calibration board includes (i) applying an adaptive thresholding algorithm the captured images to generate a binary image accentuating high-contrast calibration features, and (ii) refining detected feature locations to sub-pixel accuracy using an iterative gradient-based refinement algorithm.

[0187] Example 30. The one or more non-transitory computer-readable medium of any one of examples 21-29, wherein the computer-executable instructions further cause the one or more processors to: capture, via the optical device, a verification image of the perforated calibration board after applying the correction; analyze the verification image to determine a calibration accuracy metric; and adjust, in real time, the estimated correction parameters if the calibration accuracy metric does not meet a predetermined threshold.

[0188] Example 31. A method of calibrating a vision system, comprising: positioning a perforated calibration board at a plurality of positions relative to an optical device of the vision system; capturing, via the optical device, a plurality of images of the perforated calibration board at the plurality of positions; processing the plurality of images to analyze the perforated calibration board; establishing a correction by estimating parameters of the optical device based on the analysis of the perforated calibration board; and applying the correction to subsequent images captured by the optical device.

[0189] Example 32. The method of example 31, wherein the perforated calibration board is made of at least one of: aluminum, soda-lime glass, or PVC.

[0190] Example 33. The method of example 31 or 32, wherein the perforated calibration board has at least one bevel defining an edge of the perforated calibration board.

[0191] Example 34. The method of any one of examples 31-33, wherein the perforated calibration board has one or more perforations representative of at least one geometric shape.

[0192] Example 35. The method of any one of examples 31-34, wherein the perforated calibration board has one or more perforations representative of at least one non-geometric shape.PATENT33894 / H011

[0193] Example 36. The method of any one of examples 31 -35, wherein the perforated calibration board has at least one bevel defining one or more of the perforations.

[0194] Example 37. The method of any one of examples 31-36, wherein the optical device detects a color difference between the perforated calibration board of a first color and the at least one bevel of a second color that is different from the first color.

[0195] Example 38. The method of example 37, wherein the optical device detects the second color as darker than the first color.

[0196] Example 39. The method of any one of examples 31-38, further comprising: heating the perforated calibration board to a one or more temperatures within an operating range of the optical device, wherein the capturing comprises capturing, via the optical device, the plurality of images for each temperature of the one or more temperatures.

[0197] Example 40. The method of any one of examples 31-39, wherein the perforated calibration board at least partially defines a shadowbox for the optical device.

[0198] Example 41. The method of any one of examples 31-40, further comprising forming the perforated calibration board.

[0199] Example 42. The method of example 41, wherein the forming comprises forming one or more perforations in the perforated calibration board using a machine.

[0200] Example 43. The method of example 42, wherein the machine comprises a waterjet, laser, or milling machine.

[0201] Example 44. The method of any one of examples 31-43, wherein the optical device comprises a plurality of cameras, each camera having a sensitivity to one or more spectra, and wherein the cameras are arranged to provide an overlapping field of view.

[0202] Example 45. A method of calibrating a vision system, comprising: positioning a perforated calibration board at a plurality of positions relative to an optical device of the vision system; capturing, via the optical device, a plurality of images of the perforated calibration board at the plurality of positions, wherein the plurality of images of the perforated calibration board are analyzed to establish a correction that estimates parameters of the optical device; and applying the correction to subsequent images captured by the optical device.PATENT33894 / H011

[0203] Example 46. The method of example 45, wherein the perforated calibration board is made of at least one of: aluminum, soda-lime glass, or PVC.

[0204] Example 47. The method of example 45 or 46, wherein the perforated calibration board has at least one bevel defining an edge of the perforated calibration board.

[0205] Example 48. The method of any one of examples 45- 47, wherein the perforated calibration board has one or more perforations representative of at least one geometric shape.

[0206] Example 49. The method of any one of examples 45- 48, wherein the perforated calibration board has one or more perforations representative of at least one non-geometric shape.

[0207] Example 50. The method of any one of examples 45- 49, wherein the perforated calibration board has at least one bevel defining one or more of the perforations.

[0208] Example 51. The method of any one of examples 45- 50, wherein the optical device detects a color difference between the perforated calibration board of a first color and at least one bevel of a second color that is different from the first color.

[0209] Example 52. The method of example 51, wherein the optical device detects the second color as darker than the first color.

[0210] Example 53. The method of any one of examples 45- 52, further comprising: heating the perforated calibration board to a one or more temperatures within an operating range of the optical device, wherein the capturing comprises capturing, via the optical device, the plurality of images for each temperature of the one or more temperatures.

[0211] Example 54. The method of any one of examples 45- 53, wherein the perforated calibration board at least partially defines a shadowbox for the optical device.

[0212] Example 55. The method of any one of examples 45- 54, further comprising forming the perforated calibration board.

[0213] Example 56. The method of example 55, wherein the forming comprises forming one or more perforations in the perforated calibration board using a machine.

[0214] Example 57. The method of example 56, wherein the machine comprises a waterjet, laser, or milling machine.PATENT33894 / H011

[0215] Example 58. The method of any one of examples 45- 57, wherein the optical device comprises a plurality of cameras, each camera having a sensitivity to one or more spectra, and wherein the cameras are arranged to provide an overlapping field of view.

[0216] Example 59. A method of calibrating a vision system, comprising: positioning one of an optical device and a perforated calibration board at a plurality of positions relative to the other of the optical device and the perforated calibration board; capturing, via the optical device, a plurality of images of the perforated calibration board at the plurality of positions; processing the plurality of images to analyze the perforated calibration board; establishing a correction by estimating parameters of the optical device based on the analysis of the perforated calibration board; and applying the correction to subsequent images captured by the optical device.

[0217] Example 60. The method of example 59, further comprising: a background, wherein the background is cloth, dry wall, concrete, or a second calibration board spaced from the perforated calibration board.

[0218] Example 61. The method of example 59 or 60, wherein the perforated calibration board is made of at least one of: aluminum, soda-lime glass, or PVC.

[0219] Example 62. The method of any one of examples 59-61, wherein the perforated calibration board has at least one bevel defining an edge of the perforated calibration board.

[0220] Example 63. The method of any one of examples 59-62, wherein the perforated calibration board has one or more perforations representative of at least one geometric shape.

[0221] Example 64. The method of any one of examples 59-63, wherein the perforated calibration board has one or more perforations representative of at least one non-geometric shape.

[0222] Example 65. The method of any one of examples 59-64, wherein the perforated calibration board has at least one bevel defining one or more of the perforations.

[0223] Example 66. The method of any one of examples 59-65, wherein the optical device detects a color difference between the perforated calibration board of a first color and at least one bevel of a second color that is different from the first color.

[0224] Example 67. The method of example 66, wherein the optical device detects the second color as darker than the first color.PATENT33894 / H011

[0225] Example 68. The method of any one of examples 59-67, wherein the optical device detects at least one of (i) one or more edges defining a perforation of the perforated calibration board or (ii) one or more edges of the perforated calibration board as substantially a same color as the background.

[0226] Example 69. The method of any one of examples 59-68, wherein the optical device comprises a plurality of cameras, each camera having a sensitivity to one or more spectra, and wherein the cameras are arranged to provide an overlapping field of view.

[0227] Example 70. The method of any one of examples 59-69, further comprising: heating a second calibration board to a one or more temperatures within an operating range of the optical device; and positioning the second calibration board adjacent the perforated calibration board, wherein the capturing comprises capturing, via the optical device, the plurality of images for each temperature of the one or more temperatures.

[0228] Example 71. The method of any one of examples 59-70, wherein the perforated calibration board at least partially defines a shadowbox for the optical device.

[0229] Example 72. The method of any one of examples 59-71, further comprising forming the perforated calibration board.

[0230] Example 73. The method of example 72, wherein the forming comprises forming one or more perforations in the perforated calibration board using a machine.

[0231] Example 74. The method of example 73, wherein the machine comprises a waterjet, laser, or milling machine.

[0232] Example 75. The method of any one of examples 59-74, wherein processing the plurality of images to analyze the perforated calibration board includes associating one or more edges of at least one of (i) the one or more perforations of the perforated calibration board or (ii) the perforated calibration board with the background.

[0233] Example 76. The method of any one of examples 59-75, wherein the optical device includes at least one of (a) a near-infrared (NIR) camera, (b) a short-wave infrared (SWIR) camera, (c) a mid-wave infrared (MWIR) camera, (d) a long-wave infrared (LWIR) camera, or (e) an electro-optical (EO) camera, and wherein processing the plurality of images to analyze the perforated calibration board includes associating one or more edges of at least one of (i) the onePATENT33894 / H011 or more perforations of the perforated calibration board or (ii) the perforated calibration board with the background.

[0234] Example 77. The method of any one of examples 59-76, wherein one or more edges of at least one of (i) a perforation of the perforated calibration board or (ii) the perforated calibration board are beveled.

[0235] Example 78. A system for calibrating a vision system, comprising: a perforated calibration board; an optical device of the vision system; and one or more processors communicatively coupled to a memory storing instructions that cause the one or more processors to: position one of the optical devices and the perforated calibration board at a plurality of positions relative to the other of the optical device and the perforated calibration board; capture images via the optical device; process the images to analyze the perforated calibration board; establish a correction by estimating parameters of the optical device based on the analysis of the perforated calibration board; and apply the correction to subsequent images captured by the optical device.

[0236] Example 79. The system of example 78, further comprising a second calibration board coupled to the perforated calibration board and defining a background.

[0237] Example 80. The system of example 78 or 79, wherein the perforated calibration board is made of at least one of: aluminum, soda-lime glass, or PVC, and the background is cloth, dry wall, concrete, or a surrounding environment.

[0238] Example 81. The system of any one of examples 78-80, wherein the optical device detects one or more edges of at least one of (i) a perforation of the perforated calibration board or (ii) the perforated calibration board as substantially a same color as the background.

[0239] Example 82. The system of any one of examples 78-81, wherein processing the plurality of images to analyze the perforated calibration board includes associating one or more edges of at least one of (i) the perforation of the perforated calibration board or (ii) the perforated calibration board with the background.

[0240] Example 83. The system of any one of examples 78-82, wherein the optical device includes at least one of (a) a near-infrared (NIR) camera, (b) a short-wave infrared (SWIR) camera, (c) a mid-wave infrared (MWIR) camera, (d) a long-wave infrared (LWIR) camera, or (e) an electro-optical (EO) camera, and wherein processing the plurality of images to analyze thePATENT33894 / H011 perforated calibration board includes associating one or more edges of at least one of (i) the one or more perforations of the perforated calibration board or (ii) the perforated calibration board with the background.

[0241] Example 84. The system of any one of examples 78-83, wherein one or more edges of at least one of (i) the perforation of the perforated calibration board or (ii) the perforated calibration board are beveled.

[0242] Example 85. The system of any one of examples 78-84, wherein the optical device detects the perforated calibration board having one or more perforations representative of at least one of a geometric shape or a non-geometric shape.

[0243] Example 86. The system of any one of examples 78-85, wherein the optical device comprises a plurality of cameras, each camera having a sensitivity to one or more spectra, and wherein the cameras are arranged to provide an overlapping field of view.

[0244] Example 87. The system of any one of examples 78-86, further comprising: heating the second calibration board to one or more temperatures within an operating range of the optical device, wherein the capturing comprises capturing, via the optical device, the plurality of images for each temperature of the one or more temperatures.

[0245] Example 88. The system of any one of examples 78-87, wherein the perforated calibration board at least partially defines a shadowbox for the optical device.

[0246] Example 89. One or more non-transitory computer-readable medium having program instructions stored thereon that, when executed by one or more processors, cause the one or more processors to: position one of an optical device and a perforated calibration board at a plurality of positions relative to the other of the optical device and the perforated calibration board; capture images via the optical device; process the images to analyze the perforated calibration board; establish a correction by estimating parameters of the optical device based on the analysis of the perforated calibration board; and apply the correction to subsequent images captured by the optical device.

[0247] Example 90. The one or more non-transitory computer-readable medium of example 89, wherein the one or more processors position the perforated calibration board at the plurality of positions relative to the optical device using a robot.PATENT33894 / H011

[0248] Example 91 . The one or more non-transitory computer-readable medium of example 89 or 90, wherein the perforated calibration board is made of at least one of: aluminum, sodalime glass, or PVC and the background is cloth, dry wall, concrete, or a surrounding environment.

[0249] Example 92. The one or more non-transitory computer-readable medium of any one of examples 89-91, wherein the optical device detects one or more edges of at least one of (i) a perforation of the perforated calibration board or (ii) the perforated calibration board as substantially a same color as the background.

[0250] Example 93. The one or more non-transitory computer-readable medium of any one of examples 89-92, wherein processing the plurality of images to analyze the perforated calibration board includes associating one or more edges of at least one of (i) the perforation of the perforated calibration board or (ii) the perforated calibration board with the background.

[0251] Example 94. The one or more non-transitory computer-readable medium of any one of examples 89-93, wherein the optical device includes at least one of (a) a near-infrared (NIR) camera, (b) a short-wave infrared (SWIR) camera, (c) a mid-wave infrared (MWIR) camera, (d) a long-wave infrared (LWIR) camera, or (e) an electro-optical (EO) camera, and wherein processing the plurality of images to analyze the perforated calibration board includes associating one or more edges of at least one of (i) the one or more perforations of the perforated calibration board or (ii) the perforated calibration board with the background.

[0252] Example 95. The one or more non-transitory computer-readable medium of any one of examples 89-94, wherein one or more edges of at least one of (i) a perforation of the perforated calibration board or (ii) the perforated calibration board are beveled.

[0253] Example 96. The one or more non-transitory computer-readable medium of any one of examples 89-95, wherein the optical device detects the perforated calibration board having one or more perforations representative of at least one of a geometric shape or a non-geometric shape.

[0254] Example 97. The one or more non-transitory computer-readable medium of any one of examples 89-96, wherein the optical device comprises a plurality of cameras, each camera having a sensitivity to one or more spectra, and wherein the cameras are arranged to provide an overlapping field of view.PATENT33894 / H011

[0255] Example 98. The one or more non-transitory computer-readable medium of any one of examples 89-97, having program instructions stored thereon that further cause the one or more processors to: heat the calibration board to a one or more temperatures within an operating range of the optical device, wherein the capturing comprises capturing, via the optical device, the plurality of images for each temperature of the one or more temperatures.

[0256] Example 99. The one or more non-transitory computer-readable medium of any one of examples 89-98, wherein the perforated calibration board at least partially defines a shadowbox for the optical device.

[0257] Example 100. A fiducial for calibrating a vision system, comprising: a forward board and a rear board, each board having a front side, a back side, and a thickness defined between the front side and the back side; a spacer coupled to the forward board and the rear board, the spacer defining a distance between the boards; and a heating element coupled to the back side of the rear board and configured to heat the rear board.

[0258] Example 101. The fiducial of example 100, wherein each of the forward and rear boards is made of aluminum, soda-lime glass, or PVC.

[0259] Example 102. The fiducial of example 100 or 101, wherein the forward board has one or more perforations representative of at least one geometric or non-geometric shape.

[0260] Example 103. The fiducial of any one of examples 100-102, wherein the heating element is configured to heat the rear board to a temperature sufficient to maintain detectability in infrared (IR) for a duration exceeding an active heating period.

[0261] Example 104. The fiducial of any one of examples 100-103, wherein the front sides of the pair of boards are coated with a material configured to at least one of provide a perceived color differential in EO, reduce reflections, or ensure a uniform signature in long-wave infrared (LWIR).

[0262] Example 105. The fiducial of any one of examples 100-104, further comprising interior lighting coupled to the forward board at a position facing rearward, towards the rear board, wherein the interior lighting is thermally isolated from the forward board to minimize heat transfer.

[0263] Example 106. The fiducial of any one of examples 100-105, wherein the interior lighting comprises light-emitting diodes (LEDs).PATENT33894 / H011

[0264] Example 107. The fiducial of any one of examples 100-106, wherein the spacer comprises a plurality of spacing posts extending between the forward board and the rear board.

[0265] Example 108. The fiducial of any one of examples 100-107, wherein the spacing posts are side boards that connect each side of the forward board to the rear board, defining a shadowbox.

[0266] Example 109. The fiducial of any one of examples 100-108, wherein an inside of the shadowbox is coated with a material, to minimize heat transfer from the rear board to the forward board, wherein the material is preferably metallized polyethylene terephthalate (MPET) or Kapton.

[0267] Example 110. The fiducial of any one of examples 100-109, wherein each of the spacing posts is made of a thermally insulating material.

[0268] Example 111. The fiducial of any one of examples 100-110, wherein the thermally insulating material is garolite, PVC, polyoxymethylene, fiberglass, polyetheretherketone, nylon, or ceramic.

[0269] Example 112. The fiducial of any one of examples 100-111, further comprising a thermal control system configured to adjust the heating element to maintain a temperature difference between the forward board and the rear board.

[0270] Example 113. The fiducial of any one of examples 100-112, wherein the thermal control system further comprises one or more temperature sensors to measure a relative temperature difference between the forward board and the rear board.

[0271] Example 114. The fiducial of any one of examples 100-113, wherein each of the forward and rear boards is white on the front side and black on the back side to minimize contrast reduction.

[0272] Example 115. The fiducial of any one of examples 100-114, wherein at least one of the forward and rear boards includes a slot configured to receive a heated board.

[0273] Example 116. The fiducial of any one of examples 100-115, wherein the heated board is made of plastic.

[0274] Example 117. The fiducial of any one of examples 100-116, wherein the rear board is a calibration board with no perforations.PATENT33894 / H011

[0275] Example 118. The fiducial of any one of examples 100-117, wherein the back side of the rear board is bare aluminum.

[0276] Example 119. The fiducial of any one of examples 100-118, wherein the spacers minimize heat transfer with the one or more perforations representative of at least one geometric or non-geometric shape.

[0277] Example 120. The method of any one of examples 31-77, the system of any one of examples 78-88, and the one or more non-transitory computer-readable medium of any one of examples 89-99, wherein the perforated calibration board is part of the fiducial of examples 100- 119.

[0278] Example 121. The method of any one of examples 31-77, the system of any one of examples 78-88, and the one or more non-transitory computer-readable medium of any one of examples 89-99, wherein the plurality of cameras comprises a stereo camera and an infrared camera, wherein the stereo camera comprises two or more electro-optical cameras.

[0279] Example 122. A method of calibrating a vision system, comprising: extrinsically crosscalibrating a plurality of cameras with respect to each other camera of the plurality of cameras, wherein each camera of the plurality of cameras has a sensitivity to one or more spectra, and wherein the cross-calibration is performed regardless of the spectra to which the cameras are sensitive; determining relative positions and orientations of the plurality of cameras based on images captured by the cameras utilizing a fiducial, wherein the fiducial is positioned at a plurality of positions relative to the plurality of cameras, and images of the fiducial are captured by the plurality of cameras at the plurality of positions; and calculating a unified vision model for the vision system.

[0280] Example 123. The method of any one of examples 31-77, 120, 121, or 122, further comprising: processing imaging data of a target confirmation model having a plurality of surfaces to detect calibration features, wherein the target confirmation model is representative of a real-world target object; determining a calibration correction based on the parameters for the optical device by estimating intrinsic and extrinsic parameters from the detected features; and applying the determined calibration correction to subsequent image data captured by the optical device.PATENT33894 / H011

[0281] Example 124. The method of any one of examples 31-77, 120-123, wherein the target confirmation models have distinct thermal and / or optical characteristics.

[0282] Example 125. The method of any one of examples 31-77, 120-124, wherein the calibration features include a thermal signature of at least a portion of the target confirmation model.

[0283] Example 126. The method of any one of examples 31-77, 120-125, wherein determining the calibration correction includes entity recognition configured to associate detected calibration features with a predetermined geometric shape.

[0284] Example 127. The system of any one of examples 48-59, further comprising: a target confirmation model representative of a real-world target object; and wherein the instructions further cause the processor to: process imaging data of the target confirmation model having a plurality of surfaces to detect calibration features; determine a calibration correction based on parameters for the optical device by estimating intrinsic and extrinsic parameters from the detected features; and apply the determined calibration correction to subsequent image data captured by the optical device.

[0285] Example 128. The system of any one of examples 48-59 or 127, wherein the target confirmation model has distinct thermal and / or optical characteristics.

[0286] Example 129. The system of any one of examples 48-59, 127, or 128, wherein the calibration features include a thermal signature of at least a portion of the target confirmation model.

[0287] Example 130. The system of any one of examples 48-59 or 127-129, wherein determining the calibration correction includes entity recognition configured to associate detected calibration features with a predetermined geometric shape.

[0288] Example 131. The one or more non-transitory computer-readable medium of any one of examples 89-99 having program instructions stored thereon that further cause the one or more processors to: process imaging data of a target confirmation model having a plurality of surfaces to detect calibration features; determine a calibration correction based on parameters for the optical device by estimating intrinsic and extrinsic parameters from the detected calibration features; and apply the determined calibration correction to subsequent image data captured by the optical device.PATENT33894 / H011

[0289] Example 132. The one or more non-transitory computer-readable medium of any one of examples 89-99 or 131, wherein the target confirmation model has distinct thermal and / or optical characteristics.

[0290] Example 133. The one or more non-transitory computer-readable medium of any one of examples 89-99, 131, or 132, wherein the calibration features include a thermal signature of at least a portion of the calibration target.

[0291] Example 134. The one or more non-transitory computer-readable medium of any one of examples 89-99 or 131-133, wherein determining the calibration correction includes entity recognition configured to associate detected calibration features with a predetermined geometric shape.

[0292] Example 135. A method for confirming calibration of a vision system, comprising: processing imaging data of a target confirmation model having a plurality of surfaces to detect calibration features; determining a calibration correction based on parameters for an optical device of the vision system by estimating intrinsic and extrinsic parameters from the detected calibration features; and applying the determined calibration correction to subsequent image data captured by the optical device.

[0293] Example 136. The method of example 135, wherein the target confirmation model has distinct thermal and / or optical characteristics.

[0294] Example 137. The method of examples 135 or 136, wherein the calibration features include a thermal signature of at least a portion of the calibration target.

[0295] Example 138. The method of any one of examples 135-137, wherein determining the calibration correction includes entity recognition configured to associate detected calibration features with a predetermined geometric shape.

[0296] Example 139. A system for confirming calibration of a vision system, comprising: a target confirmation model representative of a real- world target object; a processor; and a memory storing computer-executable instructions that, when executed by the processor, cause the processor system to: process imaging data of the target confirmation model having a plurality of surfaces to detect calibration features; determine a calibration correction based on parameters for an optical device of the vision system by estimating intrinsic and extrinsic parameters fromPATENT33894 / H011 the detected features; and apply the determined calibration correction to subsequent image data captured by the optical device.

[0297] Example 140. The system of example 139, wherein the target confirmation model has distinct thermal and / or optical characteristics.

[0298] Example 141. The system of example 139 or 140, wherein the calibration features include a thermal signature of at least a portion of the calibration target.

[0299] Example 142. The system of any one of examples 139-141, wherein determining the calibration correction includes entity recognition configured to associate detected calibration features with a predetermined geometric shape.

[0300] Example 143. One or more non-transitory computer- readable medium having program instructions stored thereon that, when executed by one or more processors, cause the one or more processors to: process imaging data of a target confirmation model having a plurality of surfaces to detect calibration features; determine a calibration correction based on parameters for an optical device by estimating intrinsic and extrinsic parameters from the detected features; and apply the determined calibration correction to subsequent image data captured by the optical device.

[0301] Example 144. The one or more non-transitory computer-readable medium of example 143, wherein the target confirmation model has distinct thermal and / or optical characteristics.

[0302] Example 145. The one or more non-transitory computer-readable medium of example 143 or 144, wherein the calibration features include a thermal signature of at least a portion of the calibration target.

[0303] Example 146. The one or more non-transitory computer-readable medium of any one of examples 143-145, wherein determining the calibration correction includes entity recognition configured to associate detected calibration features with a predetermined geometric shape.

Claims

PATENT33894 / H011ClaimsWhat Is Claimed Is:

1. A method of calibrating a vision system, comprising: positioning at least one of (i) an optical device or (ii) a perforated calibration board at a plurality of positions relative to the other of (i) the optical device or (ii) the perforated calibration board; capturing, via the optical device, a plurality of images of the perforated calibration board at the plurality of positions; processing the plurality of images to analyze the perforated calibration board; establishing a correction by estimating parameters of the optical device based on the analyzed perforated calibration board; and applying the correction to subsequent images captured by the optical device.

2. The method of claim 1, further comprising: heating a second calibration board to one or more temperatures within an operating range of the optical device; and positioning the second calibration board adjacent the perforated calibration board, wherein the capturing comprises capturing, via the optical device, the plurality of images for each temperature of the one or more temperatures.

3. The method of claim 1, wherein the perforated calibration board has (i) at least one bevel defining an edge of the perforated calibration board, (ii) at least one bevel defining one or more of the perforations, or (iii) one or more perforations representative of at least one geometric or non-geometric shape.

4. The method of claim 1, wherein the optical device detects a color difference between the perforated calibration board of a first color and at least one bevel of a second color that is different from the first color, and wherein the optical device detects the second color as darker than the first color.PATENT33894 / H0115. The method of claim 1 , wherein the optical device detects at least one of (i) one or more edges defining a perforation of the perforated calibration board or (ii) one or more edges of the perforated calibration board as substantially a same color as a background.

6. The method of claim 1, wherein the optical device comprises a plurality of cameras, each camera having a sensitivity to one or more spectra, and wherein the cameras are arranged to provide an overlapping field of view.

7. The method of claim 1, wherein the optical device includes at least one of (a) a nearinfrared (NIR) camera, (b) a short-wave infrared (SWIR) camera, (c) a mid-wave infrared (MWIR) camera, (d) a long-wave infrared (LWIR) camera, or (e) an electro-optical (EO) camera, and wherein processing the plurality of images to analyze the perforated calibration board includes associating one or more edges of at least one of (i) the one or more perforations of the perforated calibration board or (ii) the perforated calibration board with a background.

8. The method of claim 1, wherein establishing the correction by estimating parameters of the optical device includes (i) detecting a predetermined pattern formed by perforations of the calibration board, (ii) comparing the predetermined pattern with a corresponding desired pattern, and (iii) correcting parameters of the optical device based on deviations between the detected predetermined pattern and the desired pattern.

9. The method of claim 1, wherein processing the plurality of images to analyze the perforated calibration board includes (i) applying an adaptive thresholding algorithm to the captured images to generate a binary image accentuating high -contrast calibration features, and (ii) refining detected feature locations to sub-pixel accuracy using an iterative gradient-based refinement algorithm.

10. The method of claim 1, further comprising: capturing, via the optical device, a verification image of the perforated calibration board after applying the correction; analyzing the verification image to determine a calibration accuracy metric; andPATENT33894 / H011 adjusting, in real time, the estimated correction parameters if the calibration accuracy metric docs not meet a predetermined threshold.

11. A system for calibrating a vision system, comprising: a perforated calibration board; an optical device of the vision system; one or more processors; and one or more memories storing computer-executable instructions that cause the one or more processors to at least: position at least one of (i) the optical devices or (ii) the perforated calibration board at a plurality of positions relative to the other of (i) the optical device or (ii) the perforated calibration board; capture, via the optical device, a plurality of images; process the images to analyze the perforated calibration board; establish a correction by estimating parameters of the optical device based on the analyzed perforated calibration board; and apply the correction to subsequent images captured by the optical device.

12. The system of claim 11, wherein the computer-executable instructions, when executed by the one or more processors, further cause the system to: heat a second calibration board to one or more temperatures within an operating range of the optical device; and position the second calibration board adjacent the perforated calibration board, wherein the capturing comprises capturing, via the optical device, the plurality of images for each temperature of the one or more temperatures.

13. The system of claim 11, wherein the perforated calibration board has (i) at least one bevel defining an edge of the perforated calibration board, (ii) at least one bevel defining one or more of the perforations, or (iii) one or more perforations representative of at least one geometric or non-geometric shape.PATENT33894 / H01114. The system of claim 11 , wherein the optical device detects a color difference between the perforated calibration board of a first color and at least one bevel of a second color that is different from the first color, and wherein the optical device detects the second color as darker than the first color.

15. The system of claim 11, wherein the optical device detects at least one of (i) one or more edges defining a perforation of the perforated calibration board or (ii) one or more edges of the perforated calibration board as substantially a same color as a background.

16. The system of claim 11, wherein the optical device comprises a plurality of cameras, each camera having a sensitivity to one or more spectra, and wherein the cameras are arranged to provide an overlapping field of view.

17. The system of claim 11, wherein the optical device includes at least one of (a) a nearinfrared (NIR) camera, (b) a short-wave infrared (SWIR) camera, (c) a mid-wave infrared (MWIR) camera, (d) a long-wave infrared (LWIR) camera, or (e) an electro-optical (EO) camera, and wherein process the plurality of images to analyze the perforated calibration board includes associating one or more edges of at least one of (i) the one or more perforations of the perforated calibration board or (ii) the perforated calibration board with a background.

18. The system of claim 11, wherein establishing the correction by estimating parameters of the optical device includes (i) detecting a predetermined pattern formed by perforations of the calibration board, (ii) comparing positions of the detected perforations with a corresponding desired pattern, and (iii) correcting parameters of the optical device based on deviations between the detected pattern and desired pattern.

19. The system of claim 11, wherein process the plurality of images to analyze the perforated calibration board includes (i) applying an adaptive thresholding algorithm the captured images to generate a binary image accentuating high -contrast calibration features, and (ii) refining detected feature locations to sub-pixel accuracy using an iterative gradient-based refinement algorithm.PATENT33894 / H01120. The system of claim 11 , wherein the computer-executable instructions, when executed by the one or more processors, further cause the system to: capture, via the optical device, a verification image of the perforated calibration board after applying the correction; analyze the verification image to determine a calibration accuracy metric; and adjust, in real time, the estimated correction parameters if the calibration accuracy metric does not meet a predetermined threshold.

21. One or more non-transitory computer-readable medium having computer-executable instructions stored thereon that, when executed by one or more processors, cause the one or more processors to at least: position at least one of (i) an optical device or (ii) a perforated calibration board at a plurality of positions relative to the other of (i) the optical device or (ii) the perforated calibration board; capture, via the optical device, a plurality of images; process the images to analyze the perforated calibration board; establish a correction by estimating parameters of the optical device based on the analyzed perforated calibration board; and apply the correction to subsequent images captured by the optical device.

22. The one or more non-transitory computer-readable medium of claim 21, wherein the computer-executable instructions further cause the one or more processors to: heat a second calibration board to one or more temperatures within an operating range of the optical device; and position the second calibration board adjacent the perforated calibration board, wherein the capturing comprises capturing, via the optical device, the plurality of images for each temperature of the one or more temperatures.

23. The one or more non-transitory computer-readable medium of claim 21, wherein the perforated calibration board has (i) at least one bevel defining an edge of the perforated calibration board, (ii) at least one bevel defining one or more of the perforations, or (iii) one or more perforations representative of at least one geometric or non-gcomctric shape.PATENT33894 / H01124. The one or more non-transitory computer-readable medium of claim 21, wherein the optical device detects a color difference between the perforated calibration board of a first color and at least one bevel of a second color that is different from the first color, and wherein the optical device detects the second color as darker than the first color.

25. The one or more non-transitory computer-readable medium of claim 21 , wherein the optical device detects at least one of (i) one or more edges defining a perforation of the perforated calibration board or (ii) one or more edges of the perforated calibration board as substantially a same color as a background.

26. The one or more non-transitory computer-readable medium of claim 21, wherein the optical device comprises a plurality of cameras, each camera having a sensitivity to one or more spectra, and wherein the cameras arc arranged to provide an overlapping field of view.

27. The one or more non-transitory computer-readable medium of claim 21, wherein the optical device includes at least one of (a) a near-infrared (NIR) camera, (b) a short-wave infrared (SWIR) camera, (c) a mid-wave infrared (MWIR) camera, (d) a long-wave infrared (LWIR) camera, or (e) an electro-optical (EO) camera, and wherein process the plurality of images to analyze the perforated calibration board includes associating one or more edges of at least one of (i) the one or more perforations of the perforated calibration board or (ii) the perforated calibration board with a background.

28. The one or more non-transitory computer-readable medium of claim 21, wherein establishing the correction by estimating parameters of the optical device includes (i) detecting a predetermined pattern formed by perforations of the calibration board, (ii) comparing positions of the detected perforations with a corresponding desired pattern, and (iii) correcting parameters of the optical device based on deviations between the detected pattern and desired pattern.

29. The one or more non-transitory computer-readable medium of claim 21, wherein process the plurality of images to analyze the perforated calibration board includes (i) applying an adaptive thresholding algorithm the captured images to generate a binary image accentuatingPATENT 33894 / H011 high-contrast calibration features, and (ii) refining detected feature locations to sub-pixel accuracy using an iterative gradient-based refinement algorithm.

30. The one or more non-transitory computer-readable medium of claim 21, wherein the computer-executable instructions further cause the one or more processors to: capture, via the optical device, a verification image of the perforated calibration board after applying the correction; analyze the verification image to determine a calibration accuracy metric; and adjust, in real time, the estimated correction parameters if the calibration accuracy metric does not meet a predetermined threshold.

31. The method of any one of claims 1 to 10, further comprising: processing imaging data of a target confirmation model having a plurality of surfaces to detect calibration features, wherein the target confirmation model is representative of a real- world target object; determining a calibration correction based on the parameters for the optical device by estimating intrinsic and extrinsic parameters from the detected features; and applying the determined calibration correction to subsequent image data captured by the optical device.

32. The method of claim 31, wherein the target confirmation models have distinct thermal and / or optical characteristics.

33. The method of claim 31 or 32, wherein the calibration features include a thermal signature of at least a portion of the target confinnation model.

34. The method of any one of claims 31 to 33, wherein determining the calibration correction includes entity recognition configured to associate detected calibration features with a predetermined geometric shape.

35. The system of any one of claims 11 to 20, further comprising:PATENT33894 / H011 a target confirmation model representative of a real-world target object; and wherein the instructions further cause the processor to: process imaging data of the target confirmation model having a plurality of surfaces to detect calibration features; determine a calibration correction based on parameters for the optical device by estimating intrinsic and extrinsic parameters from the detected features; and apply the determined calibration correction to subsequent image data captured by the optical device.

36. The system of claim 35, wherein the target confirmation model has distinct thermal and / or optical characteristics.

37. The system of claim 35 or 36, wherein the calibration features include a thermal signature of at least a portion of the target confirmation model.

38. The system of any one of claims 35 to 37, wherein determining the calibration correction includes entity recognition configured to associate detected calibration features with a predetermined geometric shape.

39. The one or more non-transitory computer-readable medium of any one of claims 21 to 30 having program instructions stored thereon that further cause the one or more processors to: process imaging data of a target confirmation model having a plurality of surfaces to detect calibration features; determine a calibration correction based on parameters for the optical device by estimating intrinsic and extrinsic parameters from the detected calibration features; and apply the determined calibration correction to subsequent image data captured by the optical device.

40. The one or more non-transitory computer-readable medium of claim 39, wherein the target confirmation model has distinct thermal and / or optical characteristics.PATENT33894 / H01141 . The one or more non-transitory computer-readable medium of claim 39 or 40, wherein the calibration features include a thermal signature of at least a portion of the calibration target.

42. The one or more non-transitory computer-readable medium of any one of claims 39 to 41, wherein determining the calibration correction includes entity recognition configured to associate detected calibration features with a predetermined geometric shape.

Images