Complex multi-vision automated cutting system

Abstract

The food processing system includes a structural frame with a conveyor that defines a processing path. The system includes a vision subsystem that captures images of opposing sides of the material on the conveyor and generates an overlaid, extracted boundary between the main product and by-product of the material based on the images. The boundary is converted into coordinates to guide an automated cutting subsystem that cuts the material along the boundary. The separation and material handling subsystem utilizes curvature in the conveyor and a shaker conveyor system to separate or provide more space between the main product and by-product after the cutting operation. The separation and material handling subsystem may also include a pick-and-place system that identifies and removes the by-product from the main product on the conveyor while also providing feedback to the vision subsystem to enhance the accuracy of the additional cutting operation.

Description

[Technical Field]

[0001] The present disclosure relates to vision-guided cutting systems that may be particularly advantageous in food processing, and more particularly, but not exclusively, to a combined multi-vision system and related methods for guiding an automated cutting system to separate main and by-products from food ingredients that may be very different in appearance. [Background technology]

[0002] Food processing has been known for many years, and one particular area within the broader industry is fish processing. Tuna is one of the world's major fisheries resources and generally can include many different species or subspecies of fish, each with different characteristics. However, tuna is a complex fish to process due to its unique characteristics relative to other fish, and as a result, most automated processing techniques are directed toward processing fish with less complex shapes, such as whitefish and salmon. As a result, tuna processing using conventional techniques remains an inefficient and expensive process.

[0003] Tuna may initially be received frozen as a whole fish. In a less conventional method of processing, whole frozen tuna is dissected into "slices" vertically along the length of the fish by cutting the frozen fish, leaving the skin, bones, dark meat, organs, and other parts (collectively, by-products) intact along with the white meat (the "main product") in the resulting frozen slices. The frozen slices then undergo additional processing techniques to separate the main product from the by-products. In most conventional applications, the whole tuna is thawed before dissecting, cleaning, and further processing into fillets, which has become the most accepted method of processing tuna, despite significant limitations and impracticalities throughout the process. As a result, methods for processing frozen slices have not been well developed to eliminate thawing from the process flow. In particular, conventional techniques for processing frozen fish present numerous challenges.

[0004] For example, some conventional methods involve knife-based cutting systems, which are generally not precise enough to efficiently and accurately process complex shapes. These problems are exacerbated when attempting to process fish of different sizes, shapes, and other characteristics, such as when processing tuna slices. Most other techniques are designed to process fish fillets instead of slices and are therefore ineffective at processing slices due to the differences between fillet and slice processing. For example, processing fish fillets requires thawing and is generally done sequentially, first removing the skin, then the bones, then the dark meat, etc., until only the main product remains. Processing slices is more complex and difficult because the number of by-products for processing in each slice is greater than the number of by-products in each successive step when processing fillets (i.e., only one by-product at a time), and each by-product in slices is smaller relative to the main product than in fillet processing. Therefore, known techniques for processing fillets are inapplicable and ineffective at processing slices.

[0005] Although some automated fish processing technologies have been proposed, these technologies also have drawbacks. For example, known automated processing technologies have insufficient vision systems, which prevent the system from obtaining the information and detail to successfully process meat slices. Known automated systems also lack the capability and flexibility to accurately and precisely cut the complex shapes and forms encountered in the industry across different sizes, shapes, and types (i.e., species or subspecies) of fish. In general, conventional automated processing technologies are inadequate for processing frozen fish slices, particularly frozen tuna slices.

[0006] As a result, it would be advantageous to have an automated food processing technique that overcomes the disadvantages and drawbacks of known systems and methods. Summary of the Invention

[0007] Generally speaking, automation in food processing benefits from intelligent, information-driven solutions. Therefore, since most raw food materials are naturally obtained, the shape and composition of each individual item is unique. To allow for a yield-efficient process, automated processing systems are preferably able to adapt to a range of different material properties and conditions.

[0008] The disclosed concepts achieve such a yield-efficient process adaptable to various material properties and conditions by combining multiple different vision technologies that collectively extract material-related information with an automated cutting system that operates based on information from the vision systems to precisely separate the primary product from the secondary product during processing. The disclosed concepts broadly include vision technologies and computational methods that combine and infer information, conveyor systems that facilitate material handling and data acquisition, controlled material flow for adjustments, and flexible cutting systems to achieve high accuracy.

[0009] More specifically, the line scan cameras of the vision subsystem image opposing sides of the material for processing, which in a non-limiting example may be tuna slices. The slices may be frozen or at least partially frozen and fully intact, meaning that they contain organs, skin, bones, and other by-products in addition to the more valuable white meat main product. This allows the vision subsystem to image opposing flat, planar sides of the tuna slices. In some examples, the vision subsystem further includes an X-ray imaging system, an infrared camera, or other imaging device to provide additional information.

[0010] Images may be captured at specific wavelengths of light and analyzed or processed at the specific wavelengths of light to determine the boundaries of different products on opposite sides of the material or slice. Information from the imaging sources can then be overlaid and interpolated to enable extraction of the product boundaries through the material, which are further converted into coordinates to guide an automated cutting subsystem. The automated cutting subsystem operates based on the coordinates to process or cut the material and separate the main product from the by-product.

[0011] The cut pieces of material are then provided to a separation and material handling subsystem, which may include a pick-and-place system to assist in separating the cut pieces and identify and remove by-products while the main product remains in the system for further processing. The pick-and-place system may be associated with an additional vision system to identify the main product and by-products and also provide a feedback loop. For example, the vision system associated with the pick-and-place system may identify that the actual result of the cut differs from the expected result of the cut relative to the extracted boundary (i.e., the final product contains additional impurities, not all by-products were accurately removed, etc.). If the difference in the finished product relative to the expected product exceeds a selected threshold, the vision system associated with the pick-and-place system instructs the vision subsystem and / or the automated cutting system to adjust the superimposed extracted boundary and / or cutting path, respectively, to eliminate the difference.

[0012] Other features and advantages of the present disclosure are provided below.

[0013] The present disclosure will be more fully understood by reference to the following drawings, in which like numerals refer to like parts throughout unless otherwise specified. The drawings do not describe every aspect of the teachings disclosed herein and do not limit the scope of the claims. [Brief explanation of the drawings]

[0014] [Figure 1] FIG. 1 is an elevation view of a tuna slice according to an embodiment of the present disclosure. [Figure 2] FIG. 1 is a cross-sectional view of a longitudinal slice of tuna according to an embodiment of the present disclosure. [Figure 3] FIG. 1 is an isometric view of a combined multi-vision automated cutting system according to an embodiment of the present disclosure. [Figure 4-1] FIG. 4 is an isometric view of the vision subsystem of the combined multi-vision automated cutting system of FIG. [Figure 4-2] 4B is a detailed schematic view of the conveyor component of the vision subsystem of FIG. 4A and a detailed schematic view of the nose bar of the conveyor component of FIG. 4B. [Figure 5] FIG. 4 is a schematic diagram of overlaid extracted boundaries from a vision subsystem for guiding an automated cutting subsystem of the combined multi-vision automated cutting system of FIG. 3. [Figure 6-1] FIG. 6A: A graphical representation showing the contrast in light intensity between different portions of a tuna at different wavelengths of light according to an embodiment of the present disclosure. [Figure 6-2] FIG. 6B: A graphical representation showing the contrast in light intensity between different portions of a tuna at different wavelengths of light according to an embodiment of the present disclosure. [Figure 6-3] FIG. 6C: A graphical representation showing the contrast in light intensity between different portions of a tuna at different wavelengths of light according to an embodiment of the present disclosure. [Figure 6-4] FIG. 6D: A series of images showing tuna slices at different peak contrast wavelengths to highlight regions of interest for processing according to embodiments of the present disclosure. [Figure 7] FIG. 4 is an isometric view of an automated cutting subsystem of the combined multi-vision automated cutting system of FIG. [Figure 8] FIG. 4 is an isometric view of the material handling and separation subsystem of the combined multi-vision automated cutting system of FIG. [Figure 9] FIG. 1 is an isometric view of a branch conveyor system for a parallel cutting operation according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0015] Those skilled in the art will appreciate that the present disclosure is illustrative only and not limiting in any way, and other embodiments of the systems and methods disclosed herein will readily suggest themselves to those skilled in the art having the aid of this disclosure.

[0016] Each of the features and teachings disclosed herein can be utilized individually or in conjunction with other features and teachings to provide automated cutting system devices, systems, and methods. Representative examples utilizing many of these additional features and teachings, individually and in combination, are described in further detail with reference to the accompanying drawings. This detailed description is intended merely to teach those skilled in the art further details for practicing aspects of the present teachings and is not intended to limit the scope of the claims. Thus, the combinations of features disclosed in the detailed description may not be necessary to practice the teachings in their broadest sense, but instead are taught merely to specifically illustrate representative examples.

[0017] Furthermore, various features of the representative examples and dependent claims may be combined in a manner not specifically and explicitly recited to provide additional useful embodiments of the present teachings. It should also be expressly noted that the designation of any range of values ​​or group of entities discloses all possible intermediate values ​​or intermediate entities for purposes of the original disclosure and for purposes of limiting the subject matter recited in the claims. It should also be expressly noted that the dimensions and shapes of the components shown in the drawings are designed to aid in understanding how the present teachings may be implemented and are not intended to limit the dimensions and shapes shown in the examples in some embodiments. In some embodiments, the dimensions and shapes of the components shown in the drawings are precisely to scale and are intended to limit the dimensions and shapes of the components.

[0018] While exemplary embodiments are described below in the context of processing frozen tuna fish, and in particular frozen slices of tuna, it should be appreciated that the disclosed concepts can be applied to any food processing technology and are not limited to processing frozen tuna. For example, the disclosed concepts are applicable to at least other food materials of flat cross-section, whether in a frozen, thawed, or cooked state. Furthermore, the disclosed concepts can be applied to any processing operation that benefits from adaptable, complex cutting patterns for the separation of different parts or portions, including, but not limited to, the processing of steaks in the meat industry, or fruits and vegetables packaged in thick slices for consumers, among others.

[0019] FIG. 1 illustrates tuna slices 20. As noted above, tuna is typically frozen whole and then broken down to produce frozen slices, such as slices 20, for further processing. The frozen slices 20 arrive at the processing facility as slices or slices of a predetermined thickness, either fully frozen or at least partially frozen. The image of slices 20 in FIG. 1 is representative of such slices typical in the industrial sector, which may be formed by slicing the tuna vertically at predetermined intervals along its length to yield slices of a selected thickness. After the frozen tuna is initially broken down into slices 20, the slices 20 are sent to a food processing facility for separation of the main product from the by-products, during which impurities are also removed. Importantly, the slices 20 typically arrive at the processing facility completely intact, meaning that viscera and other less desirable products remain in the slices 20.

[0020] For example, slices 20 in FIG. 1 include skin 22, viscera 24 in the abdominal cavity 26, dark meat 28, blood clots 30, bones 32, and white meat 34, along with other impurities and by-products not shown. Each of these features may vary in size, shape, and location, among other characteristics, in each slice 20 depending on the characteristics of the fish and the location of the slice 20. In a non-limiting example, slices 20 closer to the head of the fish may include viscera 24, whereas slices 20 closer to the tail may not include viscera 24. In a further non-limiting example, each of the above-mentioned features, i.e., skin 22, viscera 24, abdominal cavity 26, dark meat 28, blood clots 30, bones 32, and white meat 34, will generally be larger or smaller in each slice 20 depending on the characteristics of the fish from which it is derived and the location of the slice 20 in a particular fish.

[0021] FIG. 2 is a longitudinal cross-section of tuna 21 illustrating the variations in the above-mentioned characteristics of tuna 21 along its length. As shown in FIG. 2, each of the skin 22, viscera 24, belly cavity 26, dark meat 28, blood clots 30, bones 32, and white meat 34 varies in size, shape, and location, among other characteristics, along the length of tuna 21 from head 36 to tail 38 of tuna 21. Generally, white meat 34 is more desirable than other by-products, such as skin 22, viscera 24, dark meat 28, blood clots 30, bones 32, and other impurities. Due to the unique shape and characteristics of each slice 20 (FIG. 1) of tuna 21, conventional processing techniques have been unable to accurately, precisely, and efficiently process slices 20 (FIG. 1). In particular, materials of defined characteristics are highly complex for manual processing due to the number of components, their size and shape, accessibility for performing the required cuts, and the throughput at which they must be handled to produce the desired production volume. For the same reasons, it is a considerable challenge for any automated system to process such materials effectively and with the precision expected in industrial applications. As explained in more detail below, the disclosed concepts describe a processing or cutting system capable of separating desirable white meat 34 from hide 22, viscera 24, dark meat 28, blood clots 30, bones 32, and various other by-products, including impurities, while also overcoming the deficiencies and shortcomings of known systems.

[0022] FIG. 3 is an isometric view of a combined multi-vision automated cutting system 100 (sometimes referred to herein as cutting system 100, processing system, or simply system 100). System 100 includes three main subsystems, which will be described in more detail below: a vision subsystem 102, an automated cutting subsystem 104, and a material handling and separation subsystem 106. As shown in FIG. 3, each of the subsystems 102, 104, 106 is arranged sequentially (i.e., from left to right in the orientation of FIG. 3) along a processing path generally indicated by arrow A. Furthermore, the subsystems 102, 104, 106 are associated with a structural frame 108 for supporting the subsystems 102, 104, 106, and each subsystem 102, 104, 106 is aligned with the other subsystems 102, 104, 106 and each includes a respective conveyor 110 for moving material along processing path A. The conveyors 110 of each subsystem 102, 104, 106 may be associated with a corresponding drive assembly including motors, wheels, gears, chains, pulleys, belts, nose bars, etc. for driving the respective conveyors 110. Alternatively, one drive assembly may be utilized to drive all of the conveyors 110, allowing material, such as tuna slices 20, to be loaded onto the conveyor 110 and transported along processing path A, first through the vision subsystem 20, then through the automated cutting subsystem 104, and finally through the material handling and separation subsystem 106. At least some or all of the subsystems 102, 104, 106 may be in wired or wireless communication with a controller 112, indicated by dashed line 114, which may perform at least some of the functions and techniques described herein.

[0023] 4A provides additional details of the vision subsystem 102 of the combined multi-vision automated cutting system shown in FIG. 3. The vision subsystem 102 includes a structural frame 108 and a conveyor 110 configured to transport material, such as slices 20 (FIG. 1), along processing path A. In one embodiment, the conveyor 110 includes two separate and distinct conveyor components 110A, 110B separated by a space or gap 116, thereby enabling imaging of two opposite sides (i.e., above and below or top and bottom in some non-limiting examples) of the material on the conveyor sections 110A, 110B through the gap 116. The vision subsystem 102 includes at least two line scan cameras 118 and, optionally, an X-ray system 120 and one or more infrared cameras 122. The line scan cameras 118 are positioned on opposite sides of the conveyor 110 (i.e., one camera 118 is positioned on one side or above the conveyor 110 and the other camera 118 is positioned on the opposite side or below the conveyor 110) to capture two sides of the slices 20 (FIG. 1) or other material, as described above. Each line scan camera 118 is associated with a respective line light 117 (which may also be referred to as a light source 117). Additionally, the line scan cameras 118 may be mounted to a support structure that allows them to slide to adjust their position with respect to the width of the conveyor components 110A, 110B (i.e., perpendicular to the paper in the orientation of FIG. 4A), thereby assisting in the calibration and alignment of the vision subsystem 102. In the following disclosure, the combination of camera 118 and light source 117 will be referred to as a line scan camera 118 unless otherwise noted. The X-ray system 120 and one or more infrared cameras 122 may be located upstream or downstream of the line scan camera 118, or in some embodiments, may be located at (i.e., aligned with or near) the line scan camera 118.X-ray system 120 may include an X-ray source 121 and a detector 123, with source 121 located above and detector 123 located below first portion 110A of conveyor component 110. X-ray system 120 may, in some embodiments, be located upstream of line scan camera 118, with source 121 located above first conveyor component 110A to allow unobstructed passage of X-rays through objects on first conveyor component 110A to detector 123. One or more infrared cameras 122 are shown schematically by dashed lines due to the optional nature of these components and their optional placement in vision subsystem 102.

[0024] A line scan camera 118 beneath at least conveyor components 110A, 110B is positioned with a field of view that extends through or at least partially through gap 116. X-ray system 120 is for capturing images of objects within material or slices 20 (FIG. 1), which may be beneficial to processing operations but is optional depending on the particular application. Similarly, one or more infrared cameras 122 allow additional information to be captured for completeness and, in some instances, improved accuracy and precision, although the disclosed concepts are sufficiently accurate and precise for profitable production without X-ray system 120 and / or one or more infrared cameras 122. In a preferred embodiment, system 100 includes only line scan camera 118 and X-ray system 120. Because the additional information and precision gained from data acquired by X-ray system 120 is beneficial for processing tuna slices 20 (FIG. 1), one or more infrared cameras 122 are omitted to reduce costs. In some embodiments, the vision subsystem 102 can be expanded to include other imaging or measurement devices, as considered relevant to the task. Thus, the particular configuration of the vision subsystem 102 shown in FIG. 4A is non-limiting. Generally, the selection of components for the vision subsystem 102 will depend on the particular application and processing of different materials, and therefore the disclosure is not limited to only the examples provided herein.

[0025] The incident light and pixel-width exposure of the sensor or camera 118 are preferably directed at the same spot on the passing object or slice 20 ( FIG. 1 ) to enhance imaging accuracy. As a result, the gap 116 between the conveyor components 110A, 110B is preferably sized and shaped to allow a specific angular intersection between the incident light and the pixel-width exposure of the sensor or camera 118, as best shown in FIG. 4B . If the gap 116 is very narrow (i.e., less than 3 mm), a coaxial light 119 that illuminates the object parallel to the optical axis can be integrated into the vision subsystem 102. The coaxial light 119 is particularly advantageous for cameras 118 (or imaging systems) positioned below the conveyor components 110A, 110B to capture the surface of the object or slice 20 ( FIG. 1 ) from below the object in some embodiments. The cameras 118 (or imaging systems) positioned above the conveyor components 110A, 110B to capture images of the top surface of the object or slice 20 (FIG. 1) may omit coaxial lighting or may otherwise not benefit from coaxial lighting in some embodiments. In a preferred embodiment, the two imaging systems or cameras 118 are not mounted directly facing each other (i.e., not aligned with each other along a vertical axis passing through the center of each camera 118) to avoid light interference from one camera 118 to the other. Instead, the cameras 118 are preferably offset from each other by at least a few millimeters, such as 1 mm to 10 mm, or more preferably 1 mm to 5 mm. In this context only, "offset" means that a first vertical axis passing through the center of one camera 118 is not coaxial with a second vertical axis passing through the center of the other camera 118, but rather the first vertical axis is spaced apart from the second vertical axis.

[0026] The camera 118 positioned below the open gap 116 may be protected from spills or debris that may fall through the gap 116 by an air nozzle 125, which in some embodiments deflects the debris. The air nozzle 125 may be attached to a bar and / or one or more air tubes coupled to the structural frame 108 and may be configured to continuously expel air during operation, thereby directing any material or debris falling through the gap 116 away from the line-scan camera 118 below the gap 116. The nozzles 125 may be arranged in a line with equidistant spacing, or in some other arrangements, including two or more lines and irregular spacing, depending on the particular application. While two imagers or cameras 118 may be a preferred arrangement for the minimum configuration for the vision subsystem 102, the intended application for processing tuna slices 20 ( FIG. 1 ) may benefit from an X-ray imaging system 120 that provides additional, complementary information to the composition of the images from the camera 118. As mentioned above, the X-ray imaging component 120 can precede or follow the line-scan imager 118, which means that the X-ray imaging component 120 can be upstream or downstream of the camera 118 and gap 116.

[0027] FIG. 4B is a detailed view of a portion of the conveyor component 110 of the vision subsystem 102, providing additional details of the air nozzle 125 and the intersection angle of the lines of sight of the camera or image sensor 118 and the light source 117. In particular, the camera 118 may be positioned directly below the gap 116, while the light source 117 and / or the coaxial light source 119 may be positioned offset from the camera 118. The camera 118 has a line of sight or field of view indicated by dashed line 127A, and the light source 117 and / or the coaxial light source 119 output light along dashed line 127B. As shown in FIG. 4B, the field of view 127A of the camera 118 and the light 127B output by the light source 117 and / or the coaxial light source 119 are configured to intersect at the gap 116 for imaging of the cross-section (FIG. 1). The angle between dashed lines 127A, 127B may be any angle greater than or less than 15-60 degrees in some embodiments, including all intervening values ​​and limits. In some embodiments, camera 118 may be positioned just below gap 116, with light source 117 and / or coaxial light source 119 positioned at an angle relative to gap 116, as described above. As shown in FIG. 4A , line scan camera 118 above conveyor component 110 may have a different alignment and may generally be offset and upstream from line scan camera 118 below gap 116, provided that line scan camera 118 above gap 116 images the top surface of slice 20 ( FIG. 1 ) and therefore does not necessarily have a field of view through gap 116. Additionally, line scan camera 118 above conveyor component 110 may be offset from line scan camera 118 below gap 116 to avoid light interference from one camera 118 to the other. In some embodiments, both line scan cameras 118 have a field of view through gap 116, while in further embodiments, only the line scan camera 118 below gap 116 has a field of view through gap 116 to enable unobstructed imaging of the bottom surface of slice 20 (FIG. 1).

[0028] FIG. 4C is a schematic detail diagram comparing a conventional conveyor system 50 with the disclosed conveyor components 110A, 110B. In particular, the top diagram of FIG. 4C is a diagram of the conventional conveyor system 50, while the bottom diagram of FIG. 4C illustrates the conveyors 110A, 110B of the present disclosure. In the exemplary conveyor 50 shown in the top diagram, the conveyor 50 includes rollers 52 at the ends of each conveyor component 50A, 50B, which may have a diameter of approximately 3 centimeters. Thus, when two conveyors 50A, 50B are positioned with a gap 54 between them, the radii of curvature of the rollers 50 of separate conveyor segments 50A, 50B cause the object 20 to descend before the next roller 50 and conveyor 50B engages the object 20 on the other side of the gap 54, as shown schematically by dashed line 56. In other words, conventional conveyors 50 have large rounded edges associated with the relatively large diameter of rollers 50. Such an arrangement can cause object 20 to "jump," get caught in gap 54, or otherwise move on conveyors 50A, 50B at gap 54. Movement of object 20 can change the position of object 20 on conveyors 50A, 50B relative to the calibration of the vision subsystem, potentially distorting the determined boundaries of the region of interest as described herein, leading to a less accurate and less precise cutting operation.

[0029] In this manner, the disclosed conveyor components 110A, 110B are positioned to facilitate a smooth transition of material across the gap 116 while allowing imaging of objects from both sides. In some embodiments, the conveyor components 110A, 110B are nose bar or knife edge conveyors having continuously arranged nose bars (or knife edges) 129 with a diameter or radius of curvature of only a few millimeters ("mm") on each side of the gap 116. In some non-limiting examples, the nose bars 129 may have a diameter or radius of curvature of 1 mm or less, 2 mm, 3 mm, 4 mm, 5 mm, 6 mm, 7 mm, 8 mm, 9 mm, and / or 10 mm on each side. Such an arrangement creates a gap 116 large enough for line light and line scan cameras, such as camera 118 or others, to capture image data, yet narrow enough to allow objects as small as one centimeter to traverse the conveyor components 110A, 110B without significant disturbance or change in position. Thus, in some embodiments, the size of the gap 116 may be any of the dimensions described above, i.e., 1 mm or less, 2 mm, 3 mm, 4 mm, 5 mm, 6 mm, 7 mm, 8 mm, 9 mm, and / or 10 mm. While other variations are possible, as noted above, smaller gaps are preferred to avoid significant disturbance or change in the position of the material as it traverses the conveyor components 110A, 110B. Notably, the rounded edges of the conveyors 110A, 110B are significantly smaller than those of the conventional conveyor 50 because the nose bar 129 has a smaller radius of curvature than the rollers 52 of the conventional conveyor 50. In this manner, the arrangement of conveyors 110A, 110B of the present disclosure allows object 20 to traverse gap 116 or move along a straight line (nearly horizontal) without changing position, as shown by dashed line 131 in the bottom diagram of FIG. 4C, which increases accuracy and precision and avoids alignment or calibration issues due to movement of object 20 on conveyors 110A, 110B.

[0030] FIG. 5 is a schematic diagram of the overlapped extracted boundary of a slice 20 (FIG. 1) from the vision subsystem 102 used to guide the automated cutting subsystem 104, which will be described in more detail later. The overlapped extracted boundary shown in FIG. 5 is shown as a reference 124 to clarify that the estimated boundary differs from the physical slice 20 (FIG. 1) shown in FIG. 1. In most food material processing applications, it is less preferable to perform non-invasive volumetric imaging because the volumetric imaging process is a slower, less efficient process that reduces throughput and the ability to process large quantities. As a result, the disclosed concepts estimate much of the information that can actually be captured around the surface of a slice 20 or other food material and then interpolate and / or estimate any missing information through computational and model-based approximations. In operation, slices 20 (FIG. 1) are conveyed along conveyor components 110A, 110B (FIG. 4A) and camera 118 (FIG. 4A) scans or captures images of the upper and lower flat surfaces of slices 20 (FIG. 1), shown as 126A, 126B in FIG. 5, as slices 20 pass through gap 116 (FIG. 4A).

[0031] As will be described in more detail with reference to FIGS. 6A-6D , the images captured by camera 118 can be utilized, such as via controller 112, to determine the boundaries of different components of slice 20 ( FIG. 1 ). If the boundaries of the components on two opposing sides of slice 20 ( FIG. 1 ) are known, an approximation of the boundary of each component passing through slice 20 (i.e., between the two opposing sides) can be approximated by matching a line through slice 20 ( FIG. 1 ) that aligns with the identified components on the surface. In other words, the boundaries of the different components of slice 20 (i.e., dark meat 28, organs 24, etc.) are extracted from the images captured by camera 118 ( FIG. 4A ) and interpolated between the two opposing sides of slice 20 ( FIG. 1 ), thereby generating an accurate approximation of the location of each component throughout slice 20 ( FIG. 1 ), including any variations in shape or position. An illustration of such an approximation is shown in Figure 5, where all data is overlaid to form one source of reference to be analyzed, from which information is extracted for automated guidance of the automated cutting subsystem 104. By way of non-limiting example, Figure 5 shows boundary lines 128 associated with the dark meat 28 (Figure 1) on both flat surfaces 126A, 126B and through the slice 20 (Figure 1) or between the flat surfaces 126A, 126B, based on information corresponding to the location of the dark meat 28 on the flat surfaces 126A, 126B.

[0032] Figures 6A-6C are graphical representations showing the contrast in light intensity between different parts of a tuna fish at different wavelengths of light. Figure 6D is a series of images showing slices of tuna fish at different peak contrast wavelengths for highlighting regions of interest based on the concept of Figures 6A-6C. Continuing with reference to Figure 1 and beginning with Figures 6A-6C, image contrast resolution allows for differentiation of differences in the intensity of captured light, which is particularly advantageous for the system's efficiency and ability to extract desired information, depending on the application. Wavelength-specific illumination is applied to maximize detectability and identification of regions of interest in the material. For identification and separation of white meat 34, dark meat 28, and blood clot 30, images can be captured at a 737 nanometer ("nm") wavelength of light, or approximately 737 nm (i.e., 727-747 nm). For identification and separation of white meat 34 and organs 24, images can be captured at a 788 nm wavelength of light, or approximately 788 nm (i.e., 778-798 nm). For identification and separation of the white meat 34 and skin 22, images can be captured at a 1315 nanometer ("nm") wavelength of light, or approximately 1315 nm (i.e., 1305-1325 nm). While it may be possible to identify and separate the bone 32 from the remaining features of the slice 20 using light at specific wavelengths, identification of the bone 32 may particularly benefit from x-ray imaging with an x-ray system 120 (FIG. 4A) that uses both dual-energy x-rays and low-energy x-rays. Alternatively, images may all be captured at the same wavelength of light, and the image contrast resolution at the wavelengths indicated above can be utilized to distinguish differences in the captured light intensity.

[0033] For the visible light spectrum, high-frequency pulsed light is preferred to generate an alternating, intersecting line pattern that allows two different images at different wavelengths to be captured by the same line-scan camera, such as camera 118 (FIG. 4A). Additional options include the use of white illumination in sequential exposures by multiple cameras to which bandpass filters are applied for selective wavelength data. In one configuration, red illumination with a sufficiently wide distribution in the intensity spectrum can be utilized to cover both wavelengths near 737 nm and 788 nm due to the proximity of the two peaks for high contrast. Such a configuration may have particular benefits in terms of reducing cost and complexity in the imaging subsystem 102 (FIG. 4A).

[0034] The results of the above-described image capture and / or processing techniques are shown in graphical form in FIGS. 6A-6C and in photographic form in FIG. 6D. In particular, FIGS. 6A-6C represent the contrast in light intensity between features of interest at different wavelengths of light to identify peak contrast between the features at different wavelengths of light. The peak contrast generally corresponds to a preferred wavelength of light to utilize for distinguishing between different materials. Except where otherwise noted, each graph in FIGS. 6A-6C includes wavelength of light on the x-axis or horizontal axis and intensity of light on the y-axis or vertical axis. Beginning with FIG. 6A, a first graph 130A is shown providing contrast 132A between frozen white meat 134A and frozen dark meat 136A. The peak contrast in FIG. 6A is indicated by line 138A, and the peak contrast 138A between white meat 134A and dark meat 136A occurs at, or about, a wavelength of light of 737 nm. In one embodiment, the peak contrast 138A occurs specifically at 737.51 nm.

[0035] FIG. 6B provides a second graph 130B showing contrast 132B between frozen white meat 134B and viscera 136B. The peak contrast in FIG. 6B is indicated by line 138B, and the peak contrast 138B between the white meat 134B and viscera 136B occurs at, or about, a wavelength of light of 788 nm. In one embodiment, the peak contrast 138B occurs specifically at 788.39 nm. FIG. 6C provides a third graph 130C showing contrast 132C between frozen white meat 134C and skin and fat layer 136C. The peak contrast in FIG. 6C is indicated by line 138C, and the peak contrast 138C between the white meat 134C and skin and fat layer 136C occurs at, or about, a wavelength of light of 1315 nm. In one embodiment, the peak contrast 138C occurs specifically at 1315.62 nm.

[0036] FIG. 6D provides a series of three images of slice 20. These images correspond to the graphical representations in FIGS. 6A-6C and provide a visual representation of the peak contrast wavelengths. In other words, the left image in FIG. 6D corresponds to FIG. 6A, the center image in FIG. 6D corresponds to FIG. 6B, and the right image in FIG. 6D corresponds to FIG. 6C. Thus, from left to right, FIG. 6D shows images of slice 20 at 737 nm, 788 nm, and 1315 nm. Continuing to refer to FIG. 1 and beginning with the left image in FIG. 6D, the left image shows slice 20 at a peak contrast wavelength of 737 nm and a box B1 highlighting the contrast between white meat 34, dark meat 28, and blood clot 30. The center image in FIG. 6D shows slice 20 at a peak contrast wavelength of 788 nm and a box B2 highlighting the contrast between white meat 34 and viscera 24. The right image in FIG. 6D shows slice 20 at peak contrast at 1315 nm and box B3 highlighting the contrast between white meat 34, skin 22, and the fat layer 35 between skin 22 and white meat 34. As a result, FIG. 6D shows that at selected wavelengths of light, boundaries between regions of interest, such as the boundary between white meat 34 and by-products, can be more clearly identified for further processing. Additionally, the concept of FIG. 6D can be applied to slices of different shapes, sizes, and other characteristics because boundary identification is independent and unique in each slice 20, and is also faster than typical volumetric imaging techniques.

[0037] Images acquired and interpolated by the vision subsystem 102, in some cases with assistance from the controller 112 (FIG. 3), are utilized to guide the automated cutting subsystem 104 (FIG. 3). Given the integration of multiple imaging devices in the imaging subsystem 102, such as cameras 118 and others, all systems are preferably calibrated to share the same coordinate system. Returning to FIG. 4A, alignment of the vision subsystem 102 across the conveyor 110 can be achieved by one-time calibration using a visual reference object. Longitudinal synchronization can be achieved using encoders and timing of signals from optical laser triggers. Data acquired from the various imaging devices, such as at least the cameras 118, is corrected for distortions specific to the devices, allowing the image data to be overlaid for increased information.

[0038] Computational methods and machine learning models are deployed to extract information from the overlaid image data, including boundary information for different compositions of materials, using semantic model-based segmentation and extraction of key features relevant to the application. A traceable outline is derived from the processed information using an algorithm that finds pairwise correspondence sets from points with the same number of components inserted from each object's boundary above and below the object's surface, thereby forming a sequence of estimated angled trajectories to be followed as the cutting path. As further explained below, the sequence of coordinates and angles is transformed from image space (Ix, Iy) to world coordinates (Wx, Wy, Wz) and then further transformed to robot coordinates (Rx, Ry, Rz, Rq1, Rq2, Rq3, Rq4) to automate the cutting task.

[0039] Because the anticipated application for the developed system is cutting flat objects, the disclosed concepts preferably do not include three-dimensional and stereo vision cameras. World coordinates can be calculated from estimated or specified heights of materials used as intersection planes in image space. Boundary information between material components can help inform or derive the cutting path, but they are not necessarily equivalent. The cutting path has a start and end point and may follow a trajectory that deviates a predetermined distance from the determined boundary to modify or enhance the accuracy of the cut, as described in more detail below. In some embodiments, at least some or all of the above techniques are performed by the controller 112 (FIG. 3). Additional details of the controller 112 can be found in U.S. Provisional Patent Application No. 62 / 765,113, filed August 16, 2018, and International Application No. PCT / US2018 / 066314, filed December 18, 2018, both of which are incorporated herein by reference in their entireties.

[0040] In summary, the controller 112 may have a memory configured to store instructions and at least one processor configured to execute instructions to perform the above-described techniques, including, but not limited to, activating the camera 118 (FIG. 4A), acquiring image data from the camera 118 (FIG. 4A), overlaying the image data and extracting information from the overlaid image data, generating boundary information based on the overlaid image data and associated computational methods and machine learning models, generating a traceable outline based on pairwise correspondences, forming a sequence of estimated angled trajectories, transforming coordinates and angles from image space to world coordinates and from world coordinates to robot coordinates, and instructing the automated cutting subsystem 104 to perform cutting based on the robot coordinates.

[0041] FIG. 7 provides additional details of the cutting subsystem 104 of the combined multi-vision automated cutting system 100. The cutting subsystem 104 includes a structural frame 108 and a conveyor 110 for transporting slices 20 (FIG. 1) along processing path A. In one embodiment, the conveyor 110 for the cutting subsystem 104 may be different from the conveyors for the other subsystems 102, 106. In particular, the conveyor 110 for the cutting subsystem 104 may be a stainless steel chain conveyor belt with narrow links that allow small objects to be processed while providing sufficient friction to prevent the material from changing its position. In other words, the spacing between links in the conveyor 110 for the cutting subsystem 104 may be smaller than the spacing between links in the other conveyors 110, thereby increasing the contact area and friction with the material to be processed to reduce material movement on the conveyor 110 during the cutting operation in the cutting subsystem 104. Such an arrangement further aids in accuracy and precision during cutting. This is because the reduced motion ensures that the cutting subsystem 104 is guided along the correct boundary line through the material, as described above.

[0042] The cutting subsystem 140 further includes one or more cutting assemblies 140. Each cutting assembly 140 includes a cutting head 142 coupled to or associated with a respective guide assembly 144. The guide assemblies 144 may include arms 146 and links 148. The links 148 of each guide assembly 144 extend directly between one respective cutting head 142 and at least one of the arms 146 of the respective guide assembly 144. The arms 146 may be operated by an actuator or other similar drive device to move the links 148 and change the position of the cutting head 142. As shown in FIG. 7, there are six links 148 associated with the three arms 146 in each guide assembly 144 (i.e., two links 148 per arm 146). As a result, the cutting head 142 has at least six degrees of freedom, or can be moved in six different ways as a result of movement of the links 148. Alternatively, the degrees of freedom may be defined by the movement of the arms 146 to create at least six degrees of freedom for the cutting head 142. Other configurations are possible.

[0043] In one embodiment, the automated cutting subsystem 104 includes multiple cutting assemblies 140 to subdivide the tasks associated with cutting (i.e., each cutting assembly 140 may be responsible for one task of the overall cutting operation) to increase throughput and overall yield. While it may be possible to include only one cutting assembly 140 in some embodiments depending on the efficiency and speed of the cutting technique, multiple cutting assemblies 140 are preferred, and the automated cutting system 104 may include two, three, four, or more cutting assemblies 140 arranged in a line (i.e., one directly behind the other). Thus, a series of cutting assemblies 140 are tasked with performing one cutting task, such as one assembly 140 for removing the skin 22 of the slices 20 (FIG. 1), another assembly for removing the entrails 24 (FIG. 1), etc., in any selected sequence of operations related to cutting the slices 20 (FIG. 1).

[0044] In one embodiment, the cutting assembly 140 may be a waterjet cutting device with a cutting head 142 implemented as a waterjet cutting nozzle. Waterjet cutting is particularly advantageous for processing frozen food materials because it is hygienic and can process complex shapes at acceptable throughput rates, especially when multiple cutting assemblies 140 are utilized. For less demanding applications involving cutting less complex shapes, alternative cutting tool devices that are less expensive and have lower cutting accuracy may be more preferable to reduce complexity and cost when a high level of precision is not expected in the processing operation. Furthermore, waterjet cutting devices offer a high level of flexibility, precision, and accuracy, which are beneficial for certain applications processing frozen fish and / or frozen tuna, compared to other types of cutting devices.

[0045] In various embodiments, the travel speed of the robotic system, the size of the nozzle opening in the cutting head 142, and the pressure of the waterjet cutting system (i.e., cutting assembly 140) are optimized to maximize throughput while minimizing cut losses. Such characteristics of the cutting assembly 140 may also vary depending on the travel speed or throughput rate of the conveyor 110 of the cutting subsystem 104. Additionally, the characteristics of the cutting assembly 140 may vary depending on the size, type, and quantity of material being processed. In some embodiments, the water jet for each cutting assembly 140 can be turned on and off, such as by the controller 112 ( FIG. 3 ), whenever a piece of material needs to be cut. Determining the timing of the cut of each slice 20 ( FIG. 1 ), and therefore when to activate and deactivate each cutting assembly 140, may be based on sensors, such as proximity sensors, associated with the cutting assembly 140. Alternatively, such determinations may be made via the vision subsystem 102. In a non-limiting example, when slices 20 (FIG. 1) pass camera 118 (FIG. 4A) at a predetermined speed, cutting assembly 140 can be instructed to process and / or cut slices 20 (FIG. 1) based on the speed and order in which slices 20 (FIG. 1) pass camera 118, as further informed by the coordinates and overall guiding process described herein.

[0046] During operation, and after defining the cutting path from visual information (i.e., boundaries converted to coordinates with start and stop locations), an algorithm calculates the total cutting path per piece and balances the workload across multiple robotic systems or cutting assemblies 140. The robotic system may utilize full Rx-Ry-Rz translation capability to follow the material while also tracing the curve of the vision-guided cutting trajectory. Angles (Rq1, Rq2, Rq3, Rq4) are formed across the longitudinal and transverse axes of the slice 20 (FIG. 1) to best approximate the shape of the cut performed by each cutting assembly 140 or other robotic system. In this manner, the cutting subsystem 104, and more specifically the cutting assemblies 140, are guided by information from the vision subsystem 102 to automatically process material through multiple cutting assemblies 140, which are utilized to increase throughput and production.

[0047] FIG. 8 illustrates the material handling and separation subsystem 106 of the combined multi-vision automated cutting system 100 in more detail. While the automated cutting subsystem 104 physically separates different portions of the slices 20 (FIG. 1), the component portions of each slice 20 after the cutting operation remain adjacent to one another and may be in contact or close proximity to one another following the cutting operation. When processing frozen products, such an arrangement allows each separate portion after cutting to be frozen together. As a result, the system 100 includes a material handling and separation subsystem 106 to assist in the separation of each cut component, the specific separation of the main product from by-products, and the selection of a given product for removal and / or further processing, as described below.

[0048] The separation and material handling subsystem includes a structural frame 108 and a conveyor 110, as shown in FIG. 8. The separation and material handling subsystem 106 generally includes a separation section 106A and a material handling section 106B, each of which includes a respective conveyor element 110A, 110B. The separation section 106A is configured to bend each slice 20 (FIG. 1) along two major axes that are serially aligned and perpendicular to each other to aid in the initial separation of the cut components of each slice 20 (FIG. 1). The conveyor element 110A of the separation section 106A may be a sandwich conveyor belt system, including belts above and below the material being processed, both in contact with the material being processed. The conveyor element 110A includes a V- or U-shaped section 150, shown in detail view C, and a roller section 152, shown in detail view D, to create two different curvatures in the conveyor element 110A.

[0049] The first curvature formed by the V-shaped section 150 includes raised sides 154 by angled rollers 156 that support one or more belts of the conveyor element 110A. The lowest point of the curvature of each roller 156 is centrally located. Furthermore, each curved roller 156 may have the same or a different radius of curvature relative to the other rollers 156, and this radius of curvature may be constant or may vary for each roller 156. In embodiments in which each roller 156 has the same constant radius of curvature as the other rollers 156, the rollers 156 form a curvature that applies pressure across the lateral axis (i.e., left to right) of each slice 20 (FIG. 1), which creates a first break in the slice 20 (FIG. 1) along the lateral axis shown in FIG. 8.

[0050] The second curvature formed by roller section 152 includes a series of rollers positioned across the belt but at different heights, thereby forming a downward-facing arc along conveyor element 110A. In one embodiment, as shown in Detail View D, the series of rollers includes at least a first roller 158 and two second rollers 160 positioned at different heights relative to structural frame 108. First roller 158 may have a larger diameter than second rollers 160, and first roller 158 is centered with respect to second rollers 160 and positioned in the space between second rollers 160, such that second rollers 160 are positioned on each side of and spaced apart from first roller 158. The arrangement of rollers 158, 160 applies pressure to slices 20 (FIG. 1) in the longitudinal direction, thereby creating further separation between the cut components in each slice 20 (FIG. 1) along the longitudinal axis. In some applications, utilizing two curvatures with different orientations is sufficient to initially separate each slice 20 (FIG. 1), creating wider gaps between each component of the slice 20 (FIG. 1). In particular, the horizontal and vertical directions are orthogonal or perpendicular to each other, thereby creating gaps in both directions between the separated components of each slice 20 (FIG. 1).

[0051] Separation section 106A further includes a shaker conveyor system 162 downstream of conveyor element 110A to further increase the distance between each cut component of slices 20 (FIG. 1). In operation, shaker conveyor system 162 has a vibrating and / or repetitive motion to widely disperse the cut components of slices 20 (FIG. 1) while transporting the cut components along processing path A.

[0052] Material handling section 106B is downstream from separation section 106A and receives the dispersed components from shaker conveyor system 162. In one embodiment, the speed of second conveyor component 110B associated with material handling section 106B is increased relative to the speed of first conveyor component 110A to provide further separation of the cut components of rounds 20 (FIG. 1). At this stage, the cut components of rounds 20 (FIG. 1) have been dispersed sufficiently for pick-and-place system 164 to identify and pick up the by-products from second conveyor component 110B, thereby separating them from the remaining main product. In this manner, pick-and-place system 164 can transition only the main product to the next processing operation. In one embodiment, the pick and place system 164 is associated with an additional vision subsystem, which may have similarities to the vision subsystem 102, but may in some embodiments have only one line scan camera 118 above the conveyor component 110 to assess the quality of the by-product and / or main product and identify the amount of residual impurities in each individual piece in accordance with the disclosed concepts. If the amount of residual impurities exceeds a selected threshold (i.e., at least 2.5%, at least 5%, or at least 10% or more) for the calculated cutting pattern, the pick and place system 164 rejects the product from the second conveyor component 110B and / or further processing.

[0053] The characteristics and quantities of acceptable and unacceptable pieces on the second conveyor component 110B, as determined by a vision system associated with the pick-and-place system 164, may depend on the performance of the vision subsystem 102 and the accuracy of the automated cutting subsystem 104. Information collected in the separation and material handling subsystem 106 is used in some embodiments to directly adjust and control the parameterization of the computational methods utilized by the vision subsystem 102 in a feedback loop. For example, the characteristics determined by the vision system associated with the pick-and-place system 164 enable the calculation of offsets to the boundary extraction and cutting path procedures described above based on the difference between the predicted (i.e., calculated) cutting path and the actual detection results in the separation and material handling subsystem 106. As a result, the disclosed concepts contemplate computational methods in the form of self-tuning feedback loops to optimize integrated subsystems within the larger overall system 100 and balance yield loss and rejected pieces.

[0054] FIG. 9 illustrates an embodiment of a system 200 comprising an automated cutting subsystem 202 including diverging processing paths for parallel cutting operations. Depending on the characteristics of the overall system and the length and complexity of the cutting paths calculated for each round 20 (FIG. 1), bottlenecks may occur in the automated cutting subsystem. Accordingly, the system 200 includes an automated cutting subsystem 202 comprising a diverging conveyor system 204 including an extendable and / or movable belt line 206. The diverging conveyor system 204 can be provided to alternate feed into the automated cutting subsystem 202, which includes at least a first cutting subsystem 202A and a second cutting subsystem 202B arranged in parallel, by moving and / or sliding the belt line 206 along a guide bar or rail 208 extending between the entrances of the respective automated cutting subsystems 202A, 202B. Other configurations are possible, such as separate conveyors leading to each cutting subsystem 202A, 202B and divider plates that can be operated to vary the feed, among other possibilities. The bifurcated conveyor system 204 allows for increased throughput through the automated cutting subsystem 202 without significant disturbance to object positioning, thereby maintaining reference and calibration between the vision subsystem and the automated cutting subsystem. As a result, the system 200 contemplates a parallel processing arrangement for at least the automated cutting subsystem 202 to increase throughput without increasing the number of vision subsystems to maximize capacity utilization. The system 200 may include three or more cutting subsystems 202A, 202B in a similar parallel arrangement, such as at least three, four, five or more cutting subsystems arranged in parallel in the automated cutting subsystem 202.

[0055] In the above description, certain specific details are set forth to provide a thorough understanding of various embodiments of the disclosure. However, those skilled in the art will understand that the disclosure may be practiced without these specific details. In other instances, well-known structures related to the technology have not been described in detail to avoid unnecessarily obscuring the description of the embodiments of the present disclosure.

[0056] Some words and phrases used in the specification are indicated as follows: As used throughout this document, including the claims, the singular forms "a," "an," and "the" include plural references unless specifically stated otherwise. Any of the features and elements described herein may be in the singular, for example, a shell may refer to one shell. The terms "comprise" and "comprise," and their derivatives, mean inclusion without limitation. The words "associated" and "associated with," and their derivatives, may mean including contained in, interconnected, encompassing, contained in, connecting, coupled, capable of communicating, cooperating, interleaved, juxtaposed, adjacent, bounded, having, having characteristics, and the like. Other definitions of certain words and phrases are provided throughout this disclosure.

[0057] The use of ordinal numbers such as first, second, third, etc. does not necessarily imply a ranked sense of order, but may merely distinguish between multiple instances of an action or similar structure or material.

[0058] Throughout the specification, claims, and drawings, the following terms have the meaning expressly associated therewith unless the context clearly dictates otherwise. The term "herein" refers to the specification, claims, and drawings associated with this application. The phrases "in one embodiment," "in another embodiment," "various embodiments," "in some embodiments," "in other embodiments," and their derivatives refer to one or more features, structures, functions, limitations, or characteristics of the present disclosure and are not limited to the same or different embodiments unless the context explicitly dictates otherwise. As used herein, the term "or" is an inclusive "or" operator and is equivalent to the phrases "A or B, or both" or "A or B or C, or any combination thereof," with lists including additional elements being treated similarly. The term "based on" is not exclusive and allows for additional undescribed features, functions, aspects, or limitations to be based on unless the context explicitly dictates otherwise.

[0059] In general, unless otherwise specified, the materials for forming the invention and / or its components may be selected from suitable materials such as composites, ceramics, plastics, metals, polymers, thermoplastic plastics, elastomers, plastic compounds, catalysts, ammonia compounds, and the like, either alone or in any combination.

[0060] The foregoing description, for purposes of explanation, uses specific terminology and conventional phrases to provide a thorough understanding of the disclosed embodiments. It should be apparent to those skilled in the art that specific details are not required to practice the invention. The embodiments have been chosen and described to best explain the principles of the disclosed embodiments and their practical applications, thereby enabling those skilled in the art to utilize the disclosed embodiments, and various embodiments with various modifications as suited to the particular use envisioned. Therefore, the foregoing disclosure is not intended to be exhaustive or to limit the invention to the precise form disclosed, and those skilled in the art will recognize that many modifications and variations are possible in light of the above teachings.

[0061] "Top," "bottom," "upper," "lower," "up," "down," "above," "below," "left," "right," and other similar derivatives have their ordinary meaning as directional or positional designators, e.g., gravity pulls objects downward, and left refers to the direction that would be west when facing north in a cardinal orientation. These terms are not limiting with respect to the possible orientations explicitly, implicitly, or inherently disclosed in this disclosure, and unless the context clearly dictates otherwise, any aspect of the disclosed embodiments can be positioned in any orientation.

[0062] As used herein, the term "substantially" is intended to encompass normal error ranges or manufacturing tolerances due to slight variations and variations in manufacturing. Unless the context clearly dictates otherwise, relative terms such as "about," "substantially," and other derivative terms, when used to describe a value, amount, quantity, or dimension, generally refer to a value, amount, quantity, or dimension that is within plus or minus 5% of the stated value, amount, quantity, or dimension. It should be further understood that any specific dimensions of components or features provided herein are for illustrative purposes only with respect to the various embodiments described herein, and thus, the present disclosure expressly contemplates including dimensions greater or less than the stated dimension unless the context clearly dictates otherwise.

[0063] The various embodiments described above can be combined to provide further embodiments. All U.S. patents, U.S. patent application publications, U.S. patent applications, foreign applications, foreign patent applications, and non-patent publications cited herein and / or listed in the application data documents are incorporated herein by reference in their entirety. Aspects of the embodiments can be modified as necessary to incorporate concepts from the various patents, applications, and publications to provide further embodiments.

[0064] These and other modifications can be made to the embodiments in light of the above-detailed description. Generally, in the following claims, the terms used should not be construed to limit the claims to the specific embodiments disclosed in the specification and claims, but rather to include all possible embodiments, along with the full scope of equivalents to which such claims are entitled. Accordingly, the claims are not limited by the present disclosure. [Explanation of symbols]

[0065] 20 slices 21 Tuna 22 Skin 24 Internal Organs 26 Abdominal cavity 28 Dark Meat 30 Blood clot 32 bones 34 White Meat 35 Fat layer 36 Head 38 Tail 50 Conveyor System 50A, 50B Conveyor components 52 Laura 54 gap, void 56 dashed line 100 systems 102 Vision Subsystem 104 Automated Cutting Subsystem 106 Material Handling and Separation Subsystem 106A Separation Section 106B Material Handling Section 108 Structural Frame 110 Conveyor 110A, 110B Conveyor components 112 Control device 114 dashed line 116 gap, void 117 Light source 118 Line Scan Camera 119 Coaxial light 120 X-ray systems 121 X-ray source 122 Infrared Camera 123 detector 124 Standards 125 nozzles 126A, 126B plane 127A dashed line, field of view 127B dashed line, light 128 Borderline 129 Nose Bar 130A First graph 130B Second graph 130C Third graph 131 dashed line 132A, 132B, 132C contrast 134A, 134B, 134C White meat 136A Dark Meat 136B Internal Organs 136C fat layer 138A, 138B, 138C Peak Contrast 140 Cutting Assembly 142 Cutting Head 144 Guide Assembly 146 Arm 148 Links 150 V- or U-section 152 Roller Section 154 Side 156 Laura 158 First Roller 160 Second Roller 162 Shaika Conveyor System 200 systems 202 Automated Cutting Subsystem 202A First Disconnection Subsystem 202B Second Disconnection Subsystem 204 Branch Conveyor System 206 Beltline A. Processing Pathway B1, B2, B3 Boxes

Claims

1. 1. A food processing system comprising: a structural frame including a conveyor defining a processing path; a vision subsystem coupled to the structural frame and positioned along the processing path, the vision subsystem including at least two cameras configured to capture images of opposite sides of the material on the conveyor and generate overlaid extracted boundaries of primary and secondary products in the material based on the images; an automated cutting subsystem coupled to the structural frame and positioned along a processing path, the automated cutting subsystem configured to cut the material along a boundary between the main product and the by-product in the material; a separation and material handling subsystem coupled to the structural frame and positioned along the processing path, the separation and material handling subsystem configured to separate the main product and the by-product and remove the by-product from the main product; A food processing system comprising:

2. 10. The food processing system of claim 1, wherein the at least two cameras are configured to capture images of opposite sides of the ingredient at multiple different wavelengths of light to assist in generating the overlapping extracted boundaries of the main product and the by-product in the ingredient.

3. The food processing system of claim 1 , wherein the vision subsystem further comprises at least one of an X-ray imaging system and an infrared camera.

4. 10. The food processing system of claim 1, wherein the automated cutting subsystem includes at least two cutting assemblies, each cutting assembly having a cutting head with at least six degrees of freedom provided by the cutting assembly.

5. 10. The food processing system of claim 1, wherein the separation and material handling subsystem includes a shaker conveyor system configured to vibrate and separate the main product and the by-product.

6. 6. The food processing system of claim 5, wherein the separation and material handling subsystem includes a pick-and-place system configured to identify and remove the by-products from the main product along the processing path.

7. the pick and place system is associated with a vision system including at least one camera configured to capture additional images of the by-product and the main product and generate additional overlaid extracted boundaries of the main product and the by-product cut by the automated cutting subsystem; 7. The food processing system of claim 6, wherein the vision system associated with the pick-and-place system sends data to at least one of the vision subsystem and the automated cutting subsystem in a feedback loop to adjust the overlapped extracted boundaries of the main product and the by-product based on the further overlapped extracted boundaries of the main product and the by-product.

8. 1. A food processing system comprising: a structural frame including a conveyor system and a processing path along said conveyor system; a vision subsystem coupled to the structural frame and positioned along the processing path, the conveyor system including a first conveyor component and a second conveyor component associated with the vision subsystem, the first conveyor component being spaced from the second conveyor component by an air gap, the vision subsystem including at least two cameras on opposite sides of the conveyor system, one of the at least two cameras having a field of view at least partially through the air gap between the first and second conveyor components to capture an image of a bottom surface of material on the conveyor system; an automated cutting subsystem coupled to the structural frame and positioned along the processing path, the automated cutting subsystem including at least two cutting assemblies configured to cut the material along a boundary between a by-product and a main product based on information from the vision subsystem; a separation and material handling subsystem coupled to the structural frame and disposed along the processing path, the separation and material handling subsystem including a shaker conveyor system configured to separate the main product and the by-product and a pick and place system configured to identify and remove the by-product from the main product; A food processing system comprising:

9. 10. The system of claim 8, wherein the images are captured at a plurality of different wavelengths of light including a first wavelength of light between and including 727 nm and 747 nm, a second wavelength of light between and including 778 nm and 798 nm, and a third wavelength of light between and including 1305 nm and 1325 nm.

10. 9. The system of claim 8, wherein the nose bar of the first conveyor component and the nose bar of the second conveyor component each have a diameter of less than 5 mm, and the gap is less than 10 mm to transport the material across the gap from the first conveyor component to the second conveyor component in a straight line and to prevent disturbance of the position and positioning of the material.

11. 10. The system of claim 8, further comprising a plurality of air nozzles associated with one of the at least two cameras below the conveyor system, the plurality of air nozzles configured to eject air to deflect debris passing through the gap.

12. The system of claim 8 , wherein the at least two cameras are offset relative to one another.

13. 9. The system of claim 8, wherein the conveyor system includes a third conveyor component associated with the separation and material handling subsystem, the third conveyor component including at least one curvature for applying pressure to the material on the conveyor system along at least one of a lateral axis and a longitudinal axis through the material.

14. 9. The system of claim 8, wherein the conveyor system includes a branched conveyor system associated with the automated cutting subsystem, the automated cutting subsystem including at least two cutting systems arranged in parallel.

15. 1. A food processing method comprising: capturing images of opposite sides of the material at a plurality of different wavelengths of light using at least two cameras of the vision subsystem; generating overlapping extracted boundaries of the main and by-products of the material based on the image; cutting the material along the boundary between the main product and the by-product with one or more cutting assemblies of an automated cutting subassembly; separating the main product from the by-product, the separating including passing the main product and the by-product through at least one curvature of a conveyor and picking the by-product from the main product on the conveyor with a pick and place system; A food processing method comprising:

16. 16. The method of claim 15, wherein capturing images of opposite sides of the material includes positioning the at least two cameras on opposite sides of the conveyor such that a field of view of one of the at least two cameras passes at least partially through a gap between sections of the conveyor.

17. 16. The method of claim 15, wherein the plurality of different wavelengths comprises a first wavelength of about 737 nm, a second wavelength of about 788 nm, and a third wavelength of about 1315 nm.

18. 16. The method of claim 15, wherein passing the main product and the by-product through at least one curvature of the conveyor comprises passing the main product and the by-product through at least two curvatures of the conveyor and applying pressure along a transverse axis and a longitudinal axis perpendicular to the transverse axis.

19. 16. The method of claim 15, wherein separating the main product from the by-product includes vibrating the main product and the by-product by a shaker conveyor system of the conveyor.

20. 16. The method of claim 15, wherein generating the overlaid extracted boundary of the primary product and secondary product in the material comprises overlaying images of opposite sides of the material, identifying a boundary based on differences in light intensity of the primary product and secondary product at a plurality of peak contrasts associated with the plurality of different wavelengths of light, and estimating a boundary through the material based on the identified boundary on the opposite sides of the material.

Images