Method for cable insertion verification and housing inspection system

Abstract

The invention relates to a method for cable insertion verification and a housing inspection system. A method for cable insertion verification includes receiving, from a camera system, an inspection image of a connector housing held in a housing holder of a housing inspection system. The connector housing is classified as an identified connector housing type, via a connector identification machine vision system, based at least in part on a comparison between an inspection image of the connector housing and a template image corresponding to the identified connector housing type. For one or more cable cavities of the connector housing, the method includes automatically verifying whether a correct cable is inserted into the cable cavity according to the identified connector housing type.

Description

Technical Field

[0001] The present invention relates generally to computer vision systems and, more particularly, to automatic verification of cable wire insertion into connector housings. Background Art

[0002] Various types of cable connectors are commonly used to conductively couple one cable to another and / or to couple a cable to an electronic device for transmitting data and / or power. In some embodiments, such cable connectors include one or more cables that are inserted into corresponding cable cavities within a connector housing. The size and shape of the connector housing, as well as the number and distribution of cable cavities within the connector housing, can vary depending on the intended use of the cable connector. Summary of the Invention

[0003] This summary is not an extensive overview of the specification. It is not intended to identify key or critical elements of the specification, nor is it intended to delineate any unique scope of the embodiments of the specification or any scope of the claims. Its sole purpose is to present some concepts of the specification in a simplified form as a prelude to the more detailed description presented in this disclosure.

[0004] A method for cable insertion verification includes receiving an inspection image of a connector housing held in a housing holder of a housing inspection system from a camera system. Classifying the connector housing as an identified connector housing type, based at least in part on a comparison between the inspection image of the connector housing and a template image corresponding to the identified connector housing type, via a connector recognition machine vision system. For one or more cable cavities of the connector housing, the method includes automatically verifying that the correct cable is inserted into the cable cavity according to the identified connector housing type.

[0005] The features, functions, and advantages that have been discussed can be achieved independently in various embodiments or may be combined in yet other embodiments further details of which can be seen with reference to the following description and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Figure 1 An exemplary cable connector is schematically depicted, including a cable housing into which a plurality of cable wires are inserted.

[0007] Figure 2 An exemplary method for cable wire insertion verification is shown.

[0008] Figure 3A An exemplary connector housing held in a housing holder of a housing verification system is schematically depicted.

[0009] Figure 3BSchematically depicted is the use of a camera system of a housing inspection system to capture an inspection image of a connector housing.

[0010] Figure 3C An exemplary inspection image captured by a camera system is schematically depicted.

[0011] Figure 4 The connector housing type identified based on the inspection image recognition is schematically shown.

[0012] Figure 5 The corresponding relationship between the image features in the identification verification image and the template image is schematically shown.

[0013] Figure 6 Schematic diagram showing cable wire insertion verification.

[0014] Figure 7 The steps in an exemplary insertion sequence are schematically shown.

[0015] Figure 8 An exemplary computing system is schematically illustrated. DETAILED DESCRIPTION

[0016] The construction of a cable connector typically involves one or more steps in which individual cables are inserted into the cable cavity of the connector housing. This insertion can be done manually (such as by a human operator) and / or automatically (such as via a suitable mechanical or robotic insertion system). However, in either case, cable insertion can be inconsistent and prone to errors. Furthermore, manual inspection of the connector housing during manufacturing can be tedious and time-consuming, and may not always detect cable insertion errors.

[0017] Therefore, the present disclosure relates to technology for automatic cable insertion verification. The technology described herein includes capturing an inspection image of a connector housing held in a housing holder of a housing inspection system. This can be used to classify the connector housing into identified connector housing types. For example, different connector housings can have different shapes, sizes, distribution of cable cavities, etc. Therefore, by first identifying the type of connector housing being inspected, the system can confirm that the correct type of connector housing is inserted into the housing inspection system, and then evaluate whether the cable is inserted into the correct cable cavity during a subsequent cable insertion sequence. In this way, the housing inspection system can detect situations where insertion errors occur, such as, the cable is inserted into the wrong cable cavity, and / or the cable is inserted the wrong insertion distance. This can beneficially improve the speed and accuracy of constructing cable connectors, and beneficially reduce the manpower involved in such construction.

[0018] Relative to Figure 1 Schematically shows the insertion of the cable into the connector housing, Figure 1 An exemplary cable connector 100 is shown. The cable connector includes a connector housing 102 that includes a plurality of cable cavities into which cables may be inserted during assembly of the cable connector. Figure 1 The cable cavity 104 is marked in FIG. Figure 1 Three different cables 106A, 106B, and 106C are depicted. Cables 106B and 106C have been inserted into corresponding cable cavities of the connector housing, while cable 106A has not yet been inserted.

[0019] It should be understood that for the purpose of illustration, Figure 1 and others described in this article Figures 2 to 8 The specific components shown are highly simplified. Figures 1 to 8 The sizes, shapes and specific appearances of the components shown are non-limiting and are not drawn to scale. Furthermore, it should be understood that Figures 1 to 8 The components depicted in the drawings can be constructed of any suitable material. For example, as a non-limiting example, the connector housing, housing retainer, cable wires, cable contacts, and other components described herein can be constructed of any suitable combination of plastic and / or metal.

[0020] exist Figure 1 In the embodiment shown, three different cables are shown, although it will be understood that any suitable number of different cables can be inserted into the connector housing. For example, the number of cables inserted can be equal to or less than the number of cable cavities in the connector housing. In other words, it will be understood that Figure 1 The specific configuration depicted in is non-limiting, and the techniques described herein may be applied to cable connectors that connect any suitable number of cable wires to each other and / or to electronic devices such as PCBs.

[0021] The present disclosure primarily focuses on conductive cables for transmitting power and / or data. However, in some embodiments, the cable connectors described herein can be used with cables that are non-conductive but include other suitable transmission media, such as fiber optic cables.

[0022] As used herein, "cable" includes a length of material (e.g., copper wire, optical fiber) typically coated with a protective material (e.g., plastic or rubber insulation, grounded shield) used to transmit data and / or power. In other words, the term "cable" can be used to refer not only to the conductive (e.g., copper) or non-conductive (e.g., optical fiber) core of the cable, but also to any coatings, insulation, and / or shielding applied to the core.

[0023] A "cable" comprises one or more different cable wires. In the case where the cable comprises only one cable wire, the terms "cable wire" and "cable" can be used interchangeably. However, in some embodiments, a cable comprises two or more cable wires bundled together. For example, in some embodiments, the cable is a multi-conductor cable comprising two or more cable wires - for example, different conductive copper wires each covered by their own respective insulating cable jacket, and also bundled together in additional insulation and / or shielding layers to form a multi-conductor cable. In some embodiments, the cable is a shielded twisted pair cable, wherein the different cable wires comprise pairs of conductors twisted together and protected by an insulating jacket. The twisted pairs themselves are bundled together and surrounded by additional shielding and / or insulation layers to form a shielded twisted pair cable. In the case where the cable comprises two or more different cable wires, the different cable wires can each be inserted into different cable cavities of the connector housing.

[0024] Typically, there is a correspondence between different specific cables and the cable cavities into which they are inserted. For example, different specific cables may have different purposes (e.g., carrying power, carrying data, completing a ground connection) and, therefore, may be inserted into different specific cable cavities, so that a final cable connector can be used to couple the cables to the correct downstream components (e.g., ground points, input / output lines, power inlets). In some cases, different cables may have different distinguishable appearances—for example, the cables may have different sizes (e.g., gauges), may use different colors or types of insulating / protective jackets, may be made of different materials (e.g., different conductive metals or non-conductive materials) as the cable core, and / or may differ in any other suitable manner.

[0025] exist Figure 1 In some embodiments, the conductive cable contact 108 is attached to the end of the cable 106A. However, in general, the end of the cable can be processed in any suitable manner. For example, in some embodiments, a conductive contact can be attached to the end of the cable, wherein such a contact can have any suitable size and shape. In some cases, different types of conductive contacts can be attached to different cables inserted into the same connector housing. In some embodiments, the cable does not need to include a conductive contact. On the contrary, for example, the cable can terminate with a certain length of the cable core exposed, or terminate in any other suitable manner.

[0026] Each cable cavity of the connector housing is sized and shaped to accommodate a cable. As shown, cable 106B and cable 106C are inserted into corresponding cable cavities of the connector housing. The cable cavities can have any suitable size based on the size of the cable to be inserted into the cavity. In some embodiments, the same connector housing can include different cable cavity sizes to accommodate cables of different sizes (e.g., different wire gauges).

[0027] In some cases, the size of the cable cavity is designed to accommodate the insulating jacket surrounding the core of the cable (e.g., copper wire or optical fiber material) so that a length of insulated cable is inserted into the connector housing. In other embodiments, the insulating jacket can be trimmed so that only the cable core is inserted into the connector housing.

[0028] A cable of any suitable length can be inserted into the connector housing. Typically, the cable is inserted deep enough into the connector housing to enable data and / or power to be transmitted between the cable and any components coupled to the connector housing (e.g., other cables and / or electronic devices). Additionally or alternatively, the cable can be inserted deep enough so that a retaining mechanism within the connector housing holds the cable in place.

[0029] However, as discussed above, in some cases, such insertion may be prone to insertion errors—for example, the cable being inserted into the incorrect cable cavity and / or inserted the incorrect insertion distance. Manual inspection and verification of cable insertion may be tedious and time consuming. Figure 2 An exemplary method 200 for automated cable cord insertion verification is shown. The steps of method 200 may be initiated, terminated, and / or repeated at any suitable time and in response to any suitable conditions. Method 200 is primarily described as being performed by a housing inspection system that includes a controller that executes software instructions to implement a machine vision system for housing identification and insertion verification. However, it will be understood that the steps of method 200 may be performed by any suitable computing system of one or more computing devices, and any computing device that implements the steps of method 200 may have any suitable capabilities, hardware configuration, and form factor. In some embodiments, method 200 is performed by the following reference to Figure 8 The computing system 800 described is implemented.

[0030] At 202, method 200 includes receiving an inspection image of a connector housing held in a housing holder of a housing inspection system from a camera system. Figure 3A An exemplary shell inspection system is schematically shown. Specifically, Figure 3AA schematic diagram of a housing inspection system 300 for inspecting a connector housing 302 is included. The connector housing includes a plurality of cable cavities, some of which are labeled cable cavities 304. During a subsequent insertion sequence, one or more cables can be inserted into corresponding cable cavities of the connector housing, as will be described in more detail below.

[0031] The connector housing is held in a housing retainer 306. In this embodiment, the housing retainer includes two clamps that clamp the sides of the connector housing and thereby use friction to hold the connector housing in place. However, it will be understood that the housing retainer can use any suitable mechanism or structure to hold the connector housing in place while performing automatic cable insertion verification. For example, the housing retainer can use friction, suction, magnetism, adhesive and / or any other suitable force to hold the connector housing. In some embodiments, the size and shape of the housing retainer can be set to accommodate a variety of different types of housing retainers having different shapes and sizes without requiring significant reworking or reconfiguration of the housing retainer.

[0032] The connector housing can be inserted into the housing inspection system in any suitable manner. In some embodiments, an operator loads the connector housing into the housing inspection system before inserting the cable into the connector housing, and then removes the connector housing from the inspection system once the cable insertion is complete. In some embodiments, insertion and / or removal of the connector housing can be performed by a suitable automated system (e.g., a suitable automated machine or robot).

[0033] Similarly, when the connector housing is held in place by the housing retainer, the cable can be inserted into the connector housing in any suitable manner. For example, the cable can be manually inserted by an operator. Additionally or alternatively, the cable can be automatically inserted by a suitable automated system.

[0034] exist Figure 3A In an embodiment, the housing inspection system 300 is communicatively coupled to a controller 308. The "controller" takes the form of any suitable computer logic hardware configured to execute software, firmware, and / or hardware-coded instructions to thereby control the operation of the housing inspection system. For example, as described in more detail below, the controller can be used to control the operation of a camera system and / or implement a machine vision system for connector housing identification and cable insertion verification. In the event that an automated system is used to insert the connector housing into the housing inspection system and / or insert the cable into the connector housing, such automated system can be controlled by the controller 308.

[0035] In this embodiment, the controller is depicted as being separate from the shell inspection system. For example, the controller may be at least partially integrated into a structure that is physically separate from the shell inspection system and may be communicatively coupled to the shell inspection system via any suitable wired or wireless connection. However, it should be understood that in some embodiments, the controller may be "onboard" to the shell inspection system - integrated into the same physical component as the shell inspection system. In some embodiments, the controller 308 performs one or more steps of the method 200. In some embodiments, the controller 308 is implemented as described below with reference to Figure 8 The computing system 800 is described.

[0036] In some cases, before inserting the cable into the connector housing, the housing inspection system receives a set of cable insertion parameters including information related to the cable insertion process. Figure 3A In the embodiment of FIG, the controller 308 receives a set of cable insertion parameters 310. Generally, these include any information relevant to the subsequent cable insertion process, wherein the cable wire is inserted into the cable cavity of the connector housing.

[0037] For example, in some cases, the cable insertion parameters specify the expected type of connector housing that should be inserted into the housing inspection system for the current assembly process. Once the connector housing is inserted, the housing inspection system can perform connector housing identification to confirm that the connector housing is of the correct type. Additionally or alternatively, the cable insertion parameters can specify the correct insertion sequence for one or more cables into one or more cable cavities of the connector housing. This can include the order in which the cable cavities are filled (e.g., based on different identifiers assigned to the cable cavities), the order in which different cables are inserted (e.g., based on different identifiers assigned to the cables), the mapping of different specific cables to different cable cavities, and / or any other suitable information. It should be understood that the cable insertion parameters can include any suitable information related to the cables inserted into the connector housing.

[0038] The cable insertion parameters can be provided in any suitable manner. In some embodiments, the cable insertion parameters are provided by a human user, such as an operator or supervisor of the shell inspection system. For example, the human user may specify the cable insertion parameters by providing input to a suitable input mechanism, such as a computer mouse, keyboard, and / or touch-sensitive display interface. Additionally or alternatively, the cable insertion parameters may be retrieved from computer memory. For example, the cable insertion parameters may be stored in local data storage hardware of the shell inspection system, loaded from a removable storage device, and / or accessed via a computer network.

[0039] In some embodiments, various aspects of the cable insertion parameters can be displayed for review. For example, the cable insertion parameters can include a set of instructions for a human user to insert the cable wires in the correct order. In some embodiments, the displayed instructions are updated in real time, for example, to provide feedback to the user regarding whether the correct cable wires were inserted in the most recent insertion step.

[0040] Figure 3B Schematically provided are different views of a connector housing held by a housing holder of a housing inspection system. Figure 3B In the embodiment, portions of the housing inspection system are omitted to provide a top view of the connector housing 302 (e.g., along the Figure 3A and Figure 3B 312. The connector housing is held within the housing inspection system 300 by the housing holder 306. From this perspective, the housing inspection system can be seen to include a camera system 312. The camera system is configured to capture an inspection image 314 of the connector housing, as will be described in greater detail below.

[0041] exist Figure 3B In an embodiment, the camera system 312 is a stereo camera system including a first camera 316A and a second camera 316B. Each of these cameras can capture its own corresponding inspection image of the connector housing, and one or both of these inspection images can be used for connector housing identification and cable insertion verification. However, typically, the camera system includes any suitable number and type of different cameras for capturing insertion images. For example, in some embodiments, the camera system may include only a single camera. In some embodiments, the camera system may include three or more cameras. Typically, increasing the number of cameras can improve the accuracy of the connector identification and insertion verification process, but increases the complexity and expense of the housing inspection system. In other words, the housing inspection system includes a camera system having one or more suitable cameras, wherein each camera can be sensitive to any suitable wavelength of electromagnetic radiation and have any suitable image capture capabilities, including resolution, frame rate and / or field of view.

[0042] As one embodiment, the camera system includes one or more grayscale cameras or RGB cameras that are sensitive to visible wavelengths of light and output grayscale images or RGB images. In some embodiments, the camera system includes one or more depth cameras in addition to or in place of visible light cameras and / or other suitable cameras. The depth cameras are configured to output depth images, wherein pixels of the depth images encode the detected distances between the depth camera's image sensor and physical objects in the surrounding environment. Any suitable depth sensing technology can be used—for example, stereoscopic, structured light, or time-of-flight.

[0043] In embodiments where both a visible light camera and a depth camera are used, they can in some cases be used together as an integrated camera module. As a non-limiting example, an Intel® RealSense™ camera system can be used, which includes an RGB camera module and a depth camera module that are both in known alignment and output both RGB image data and depth image data.

[0044] exist Figure 3B In an embodiment, the housing inspection system further includes an illumination system 318. The illumination system is configured to emit illumination light toward the connector housing. This can be used to provide relatively uniform illumination conditions when capturing inspection images of the connector housing. The illumination system can take any suitable form and can use any suitable hardware components to generate the illumination light. The illumination light can have any suitable intensity and use any suitable wavelength of electromagnetic radiation.

[0045] The illumination light may be emitted at any suitable time. For example, in some cases, the illumination light may be provided whenever the housing inspection system is powered on. In some cases, the illumination light may be selectively turned on and off. For example, the illumination light may be emitted only when the connector housing is held in the housing holder, or may be emitted only immediately before capturing an inspection image (e.g., the illumination light may function as a camera flash).

[0046] exist Figure 3B In an embodiment, the camera system is attached to the housing inspection system at a fixed position relative to the housing holder. This can beneficially improve the consistency of the inspection images captured by the camera system. For example, because the position of the camera system is fixed relative to the housing holder, different connector housings held in the housing holder can be imaged from approximately the same distance for each inspection image. In some cases, the housing holder and / or the connector housing are designed so that when the connector housing is held in the housing holder, the connector housing is positioned at a known fixed distance away from the camera system. For example, the housing holder can be designed to fit within a notch or groove in the connector housing. In some cases, when the connector housing is loaded into the housing holder, the connector housing is inserted so that features or markings on the connector housing are aligned with features or markings on the housing holder, thereby indicating that the connector housing is properly positioned relative to the camera system.

[0047] As shown in the figure, in this embodiment, the connector housing has two different faces on opposite sides of the connector housing (relative to Figure 3A and Figure 3B This includes the insertion side where the cables are inserted and the viewing side that faces the camera system. Figure 3BIn the example, connector housing 302 includes an insertion surface 320 and a viewing surface 322. The cable cavity of the connector housing extends from the insertion surface of the connector housing to the viewing surface of the connector housing. Furthermore, the inspection image captured by the camera system depicts the viewing surface. In this way, when the cable is inserted into the connector housing, the end of the cable can be made visible in the inspection image captured by the camera system. This can be used to automatically verify correct cable insertion, as will be described in more detail below.

[0048] Figure 3C A schematic diagram includes an exemplary inspection image 314 of connector housing 302 captured by camera system 312. As described above, inspection image 314 depicts the viewing side of the connector housing, which is opposite the insertion side (in which the cable is inserted). In this embodiment, the cable lumen extends from the insertion side through the connector housing to the viewing side. Therefore, any cable inserted into the connector housing is visible in the inspection image captured from the viewing side.

[0049] It will be understood that the images described herein need not be visually rendered or displayed for viewing by a human user. Figure 3C An exemplary inspection image is shown in , but this is for illustration purposes only. Rather, in some embodiments, the image is captured and stored by the shell inspection system for processing into a digital data structure that is never visually represented on a computer display, or otherwise presented for viewing.

[0050] Brief return Figure 2 At 204, method 200 includes classifying the connector housing into the identified connector housing type using the connector identification machine vision system. This can be performed based at least in part on a comparison between the inspection image of the connector housing and a template image corresponding to the identified connector housing type.

[0051] Relative to Figure 4 This process is shown schematically. Figure 4 An exemplary controller 400 is shown. Figure 3A Like the controller 308, the controller 400 can be implemented as any suitable computer logic device. In some embodiments, the controller 400 is implemented as Figure 8 The computing system 800 is described. Figure 4 In FIG. 4 , a controller 400 is communicatively coupled to a camera system 402 of a housing inspection system. The controller 400 receives an inspection image 404 depicting a connector housing held in a housing holder from the camera system 402, such as a Figure 3C The inspection image 314 is input to the connector identification machine vision system 406, which is configured to classify the connector housing in the inspection image into an identified connector housing type.

[0052] The connector recognition machine vision system is implemented in any suitable manner. In some embodiments, the connector recognition machine vision system includes one or more suitable artificial intelligence (AI) and / or machine learning (ML) models configured to classify input images. As a non-limiting example, the connector recognition machine vision system can be trained with a plurality of different template images corresponding to a plurality of different connector housing types. Figure 4 4, wherein a connector recognition machine vision system is trained with a plurality of different template images including template images 408A-408C. Based on an inspection image 404 obtained from a camera system 402 and the template images 408A-408C, the connector recognition machine vision system 406 outputs an identified connector housing type 410, e.g., classifying a connector housing depicted in the inspection image as one of the connector housing types represented in the template images.

[0053] The connector recognition machine vision system can be implemented using any suitable ML and / or AI technology. In some cases, the connector recognition machine vision system includes a support vector machine that can be used to generate different classification models corresponding to different recognized connector housing types. Additionally or alternatively, the connector recognition machine vision system can include an artificial neural network. For example, the connector recognition machine vision system can include one of a support vector machine and an artificial neural network. Figure 3B In some cases, the connector recognition machine vision system may include a first recognition model trained to classify inspection images output by a first camera of a stereo camera system and a second recognition model trained to classify inspection images output by a second camera of the stereo camera system. Thus, in one non-limiting manner, training the connector recognition machine vision system may include placing a connector housing in a housing inspection system and then capturing one or more images of the connector housing (referred to as “template” images). This may be repeated for each connector housing type. These template images may then be loaded into a database where an image classification model is generated (e.g., using a support vector machine or artificial neural network) to generate separate and unique image features for each connector type. In some embodiments, the “template image” may include a three-dimensional model or scan in addition to or in place of a two-dimensional image of the connector housing.

[0054] In any case, the connector identification machine vision system is configured to output an identified connector housing type based at least in part on the inspection image. In some embodiments, the classification of the connector housing includes confirming that the correct type of connector housing is inserted into the housing inspection system. For example, in some cases, the "correct" type of connector housing is known, such as may be described above with respect to Figure 3A The cable insertion parameters discussed above are specified. Therefore, classifying the connector housings can be used to confirm that the correct type of housing is inserted. In some cases, if an incorrect connector housing type is detected, the housing verification system can output an error. This can, for example, detect a situation where an operator mistakenly loads the wrong type of connector housing into the housing verification system.

[0055] Alternatively, in some embodiments, the connector housings are classified generally blindly, e.g., without a priori knowledge of the "correct" connector housing type for the current type of cable connector being constructed. In one embodiment, a user can load a suitable connector housing into a housing inspection system, which can then classify the connector housing as the identified connector housing type. From there, the system can automatically retrieve the correct insertion sequence and insertion instruction set for the identified connector housing type, which the user can then follow when inserting a cable into the connector housing.

[0056] The output of the connector identification machine vision system can take any suitable form. As a non-limiting example, the inspection image can be provided to different classification models corresponding to different identified connector housing types, and each classification model then outputs a prediction score indicating the confidence of the model that the inspection image corresponds to the connector housing type of the model. These prediction scores can then be aggregated and compared to a classification threshold. If the most likely prediction score exceeds the classification threshold, the machine vision system classifies the connector housing as belonging to the connector housing type corresponding to the most likely prediction score. If no prediction score exceeds the classification threshold, the machine vision system can output a classification error. Depending on the implementation, the classification threshold can have any suitable value, for example, it can be set by a user or operator to balance the risk of misclassification with the risk of misclassification.

[0057] Once the connector housing is classified, the housing inspection system may, in some cases, perform cavity detection to identify the location of the connector housing's cable cavity. In some embodiments, detecting the location of the cable cavity may include detecting a correspondence between image features in the inspection image and image features in the template image. For example, once a template image of a particular connector housing type is captured, the location of the cable cavity within the template image may be manually marked by a human user and / or automatically detected via a suitable computer vision system (e.g., a suitable ML and / or AI model). Thus, by detecting the correspondence between the inspection image and the template image, the location of the cable cavity within the inspection image can be detected.

[0058] This is relative to Figure 5 Schematically shown, an exemplary inspection image 500 is included depicting a connector housing held in a housing inspection system. The inspection image 500 is compared to a template image 502 depicting the same type of connector housing as the inspection image. Figure 5 , image features 504A and 504B of the test image 500 correspond to image features 506A and 506B of the template image 502 .

[0059] Any suitable method can be used to detect correspondences between such image features. In some cases, a suitable feature recognition algorithm such as SIFT (Scale Invariant Feature Transform) and / or FLANN (Fast Approximate Nearest Neighbor Search) can be used to identify correspondences between image features in two different images. In one exemplary method, SIFT can be initially applied to the two images to detect keypoints and compute their descriptors. SIFT keypoints are points in the image that are scale and rotation invariant, and each keypoint has an associated descriptor. These descriptors effectively capture local gradient information around the keypoint, making them discriminative and robust to viewpoint and illumination changes.

[0060] Once SIFT descriptors are obtained from both images, corresponding descriptors between the two images can be matched using FLANN. FLANN is an algorithm for efficiently finding approximate nearest neighbors in a high-dimensional space. It accelerates the search for the closest match between each descriptor in one image and a descriptor in the other image. During this process, each descriptor in the first image is compared to all descriptors in the second image to find the best match. The FLANN algorithm effectively searches for the nearest neighbor (i.e., the most similar descriptor) in this high-dimensional space. Typically, the nearest neighbor is identified as the best match, but sometimes the second-nearest neighbor is also considered to ensure uniqueness of the match and filter out false matches. The output of this process is a set of keypoint pairs, where each keypoint pair consists of a keypoint from the first image and its corresponding keypoint in the second image. For example, these keypoints could represent matching cable cavities between the two images.

[0061] In some embodiments, detecting correspondence between the cable cavities in the inspection image and the template image includes generating a homography matrix to account for rotation of the position of one or more cable cavities caused by rotation of the connector housing in the inspection image relative to the template image. Figure 5 This is schematically shown where a homography matrix 508 is generated to account for the rotation of the connector housing between the template image and the verification image.

[0062] In general, the homography matrix can be generated in any suitable manner. In the case of SIFT, because SIFT descriptors are rotation-invariant, the matched keypoints will naturally account for any rotation of the connector housing between the two images. This means that even if the connector housing is rotated in one image relative to the other, the SIFT algorithm will still be able to find corresponding keypoints. Using the matched keypoints, a homography matrix can be estimated, which describes how points in one image are transformed to points in the other image under a planar perspective transformation. This includes translation, rotation, scaling, and perspective distortion. The goal is to solve for a homography matrix H that satisfies the equation p'=Hp, where p and p' are corresponding points in the two images. This can be accomplished using methods such as direct linear transformation (DLT) or using algorithms such as random sample consensus (RANSAC) to robustly handle outliers. The result is that when applied to points in one image, the homography matrix maps them to their corresponding points in the other image. This matrix accounts for all planar transformations, including any rotation that occurs between the two views of the object.

[0063] Brief return Figure 2 At 206, method 200 includes automatically verifying whether the correct cable is inserted into the cable cavity for one or more cable cavities of the connector housing according to the identified connector housing type. Typically, a group of cables can be inserted into the one or more cable cavities according to a series of one or more cable insertion steps of an insertion sequence. After each insertion step, and / or after the entire insertion sequence is completed, the images captured by the camera sequence can be used to verify whether the cables are correctly inserted, for example, the correct cable is inserted into the correct corresponding cable cavity. In other words, in some embodiments, the housing inspection system can verify whether the most recently inserted cable is inserted into the correct cable cavity in the one or more cable cavities after each insertion step. Additionally or alternatively, each cable can be verified immediately after completing the entire insertion sequence.

[0064] Relative to Figure 6 An example method of cable wire insertion verification is schematically illustrated. Figure 6An exemplary image 600 of a connector housing 601 is depicted. The connector housing 601 includes a plurality of cable cavities, one of which is labeled cable cavity 602. In this embodiment, all of the cable cavities of the connector housing 601 are currently empty, but this need not always be the case. Instead, relative to Figure 6 The illustrated method may be performed any time a connector housing includes at least one cable cavity that is still empty.Image 600 is referred to as a "pre-insertion image" because it is captured prior to the connector insertion step 604 of the cable insertion sequence.

[0065] After the connector insertion step 604, a post-insertion image 606 is captured using the camera system, again depicting the connector housing 601. However, in this case, the cable 608 has already been inserted into the cable cavity 602, and the cable is partially visible in the post-insertion image 606. In this exemplary embodiment, both the pre-insertion image 600 and the post-insertion image 606 are input into an image subtraction process 610, which subtracts the pixel values of the post-insertion image from the pixel values of the pre-insertion image (or vice versa). Because the only change between the two images is the insertion of the cable, many pixel values in the image subtraction result 612 will be at or near zero. Therefore, the image subtraction result will have the effect of highlighting any changes in the appearance of the connector housing between the two captured images, such as the pixels depicting the cable 608 inserted into the cable cavity 602. In other words, the housing inspection system generates an image subtraction result based on the pre-insertion and post-insertion images, which can be used to detect movement (e.g., the insertion of a cable) between the two images.

[0066] exist Figure 6In the method, image subtraction results 612 are provided to an insertion verification machine vision system 614, which is trained to output a probability of correct insertion 616 based on the image subtraction results. In some cases, when motion is detected in the image subtraction results (e.g., a region of pixels having pixel values above a motion threshold), a region of interest centered around the detected motion is input to the insertion verification machine vision system 614. If the probability of correct insertion 616 exceeds the verification threshold, the housing inspection system can verify that the correct cable has been inserted. If the probability of correct insertion 616 does not exceed the verification threshold, the housing inspection system can output an error indicating that an incorrect insertion has occurred. This can, for example, prompt an operator to remove the most recently inserted cable and try again. It will be appreciated that the probability threshold can have any suitable value, depending on the implementation, for example, to balance the risk of false error notifications with the risk of undetected insertion errors. In some cases, this process can be performed for each image captured by the camera system. For example, if the camera system is a stereo camera system that simultaneously captures two different images, each of these images can be used as the basis for performing cable insertion verification.

[0067] The insertion verification machine vision system can take any suitable form and can be trained in any suitable manner. In one exemplary embodiment, training the insertion verification machine vision system can include first capturing one or more images of an empty connector housing (e.g., an "empty image"). Next, after the cable cavity of the connector housing is filled with the correct corresponding cable wire, additional images of the connector housing can be captured (e.g., a "filled image"). These empty and filled reference images can then be used to train a classification algorithm, which can be used to infer whether the cable wire is correctly inserted into the connector housing. Similar to the connector recognition machine vision system, any suitable ML and / or AI technology can be used. As a non-limiting example, the insertion verification machine learning system can use a support vector machine.

[0068] In some cases, insertion of the "correct" cable may include determining that the correct cable cavity is filled according to the insertion sequence. In other words, in some cases, the system does not distinguish between different specific cables, but rather assumes that if a cable is inserted into the next cable cavity in the insertion sequence, the sequence has been correctly followed. Alternatively, in some cases, where different cables have sufficiently different appearances (e.g., size, color, contact type), cable insertion verification may include determining that the correct individual cables are inserted into the correct corresponding cable cavity in the connector housing.

[0069] in any case, Figure 6The illustrated process can be repeated any suitable number of times throughout the insertion sequence, wherein one or more different cables are inserted into the cable cavity of the connector housing. As described above, in some cases, insertion verification is performed after each insertion step in the insertion sequence. In other embodiments, insertion verification may be performed only after the entire insertion sequence is completed (e.g., after each cable has been inserted).

[0070] Figure 7 An exemplary insertion sequence 700 is schematically illustrated. As shown, in this embodiment, the insertion sequence includes a plurality of steps 702A and 702B, wherein different cable wires are inserted into the connector housing. Figure 7 In the embodiment of the present invention, after each insertion step, the housing inspection system performs an insertion verification step 704A / 704B to determine whether the most recent insertion step was correct, for example, the correct cable was inserted into the correct corresponding cable cavity. Depending on the number of cable cavities in the connector housing and the number of cables to be inserted, the insertion sequence 700 can continue with any suitable number of subsequent steps.

[0071] The methods and processes described herein may be bound to a computing system of one or more computing devices. Specifically, such methods and processes may be implemented as an executable computer application, a network-accessible computing service, an application programming interface (API), a library, or a combination of the above and / or other computing resources.

[0072] Figure 8 Schematically illustrates a simplified representation of a computing system 800 configured to provide any of the computing functions described herein. Computing system 800 may take the form of one or more network accessible devices, personal computers, server computers, mobile computing devices, and / or other computing devices.

[0073] The computing system 800 includes a logic subsystem 802 and a storage subsystem 804. The computing system 800 may optionally include a display subsystem 806, an input subsystem 808, a communication subsystem 810, and / or Figure 8 Other subsystems not shown.

[0074] The logic subsystem 802 includes one or more physical devices configured to execute instructions. For example, the logic subsystem can be configured to execute instructions as part of one or more applications, services, or other logical constructs. The logic subsystem can include one or more hardware processors configured to execute software instructions. Additionally or alternatively, the logic subsystem can include one or more hardware or firmware devices configured to execute hardware or firmware instructions. The processors of the logic subsystem can be single-core or multi-core, and the instructions executed thereon can be configured for sequential, parallel, and / or distributed processing. The components of the logic subsystem can optionally be distributed across two or more separate devices that can be remotely located and / or configured for collaborative processing. Aspects of the logic subsystem can be virtualized and executed by a remotely accessible, networked computing device configured in a cloud computing configuration.

[0075] The storage subsystem 804 includes one or more physical devices configured to temporarily and / or permanently store computer information, such as data and instructions executed by the logic subsystem. When the storage subsystem includes two or more devices, the devices may be collocated and / or remotely located. The storage subsystem 804 may include volatile, non-volatile, dynamic, static, read / write, read-only, random access, sequential access, location addressable, file addressable, and / or content addressable devices. The storage subsystem 804 may include removable and / or internal devices. As the logic subsystem executes instructions, the state of the storage subsystem 804 may transition, for example, to store different data.

[0076] Aspects of the logic subsystem 802 and the storage subsystem 804 may be integrated together into one or more hardware logic components. Such hardware logic components may include, for example, program and application specific integrated circuits (PASIC / ASIC), program and application specific standard products (PSSP / ASSP), systems on chips (SOCs), and complex programmable logic devices (CPLDs).

[0077] The logic subsystem and the storage subsystem can collaborate to instantiate one or more logical machines. As used herein, the term "machine" is used to collectively refer to a combination of hardware, firmware, software, instructions, and / or any other components that collaborate to provide computer functionality. In other words, a "machine" is never an abstract concept but always has a concrete form. A machine can be instantiated by a single computing device, or a machine can include two or more subcomponents instantiated by two or more different computing devices. In some implementations, a machine includes a local component (e.g., a software application executed by a computer processor) that collaborates with a remote component (e.g., a cloud computing service provided by a server computer network). The software and / or other instructions that give a particular machine its functionality can optionally be stored as one or more unexecuted modules on one or more suitable storage devices.

[0078] When a display subsystem is included, the display subsystem 806 can be used to present a visual representation of the data held by the storage subsystem 804. This visual representation can take the form of a graphical user interface (GUI). The display subsystem 806 can include one or more display devices utilizing virtually any type of technology. In some implementations, the display subsystem can include one or more virtual, augmented, or mixed reality displays.

[0079] When an input subsystem is included, the input subsystem 808 may include or interface with one or more input devices. The input device may include a sensor device or a user input device. Examples of user input devices include a keyboard, a mouse, a touch screen, or a game controller. In some embodiments, the input subsystem may include or interface with selected natural user input (NUI) components. Such components may be integrated or peripheral, and the conversion and / or processing of input actions may be handled on-board or off-board. Exemplary NUI components may include microphones for voice and / or sound recognition; infrared, color, stereo, and / or depth cameras for machine vision and / or gesture recognition; and head trackers, eye trackers, accelerometers, and / or gyroscopes for motion detection and / or intent recognition.

[0080] When a communication subsystem is included, the communication subsystem 810 can be configured to communicatively couple the computing system 800 with one or more other computing devices. The communication subsystem 810 can include wired and / or wireless communication devices compatible with one or more different communication protocols. The communication subsystem can be configured to communicate via personal, local, and / or wide area networks.

[0081] The present disclosure is presented by way of example and with reference to the associated drawings. Components, processing steps, and other elements that may be substantially the same in one or more of the drawings are co-identified and described with minimal repetition. However, it will be noted that the co-identified elements may also differ to some extent. It will also be noted that some of the figures may be schematic and not drawn to scale. The various drawing scales, aspect ratios, and numbers of parts shown in the drawings may be varied to make certain features or relationships easier to see.

[0082] In an embodiment, a method for cable insertion verification includes: receiving an inspection image of a connector housing held in a housing holder of a housing inspection system from a camera system; classifying the connector housing as an identified connector housing type via a connector recognition machine vision system based at least in part on a comparison between the inspection image of the connector housing and a template image corresponding to the identified connector housing type; and automatically verifying, for one or more cable cavities of the connector housing, whether a correct cable is inserted into the cable cavity according to the identified connector housing type. In this or any other embodiment, the connector recognition machine vision system is trained with a plurality of different template images corresponding to a plurality of different connector housing types. In this or any other embodiment, the connector recognition machine vision system includes one of a support vector machine and an artificial neural network. In this or any other embodiment, the camera system is a stereo camera system including a first camera and a second camera, and wherein the connector recognition machine vision system includes a first recognition model trained to classify the inspection image output by the first camera and a second recognition model trained to classify the inspection image output by the second camera. In this or any other embodiment, the camera system is attached to the housing inspection system at a fixed position relative to the housing holder. In this or any other embodiment, one or more cable cavities of the connector housing extend from an insertion face of the connector housing to a viewing face of the connector housing, and wherein the inspection image of the connector housing depicts the viewing face. In this or any other embodiment, the housing inspection system further includes an illumination system configured to emit illumination light toward the viewing face of the connector housing. In this or any other embodiment, the method further includes detecting the position of the one or more cable cavities in the inspection image by detecting a correspondence between image features in the inspection image and image features in a template image. In this or any other embodiment, the method further includes generating a homography matrix to account for rotation of the position of the one or more cable cavities caused by rotation of the connector housing relative to the template image. In this or any other embodiment, a group of cables are inserted into the one or more cable cavities according to a series of one or more cable insertion steps in an insertion sequence, and wherein, after each insertion step, the method further includes automatically verifying that the most recently inserted cable is inserted into the correct one of the one or more cable cavities. In this or any other embodiment, the method further includes capturing a pre-insertion image before each insertion step, capturing a post-insertion image after each insertion step, and generating an image subtraction result based on the pre-insertion image and the post-insertion image. In this or any other embodiment, the image subtraction results are provided to an insertion verification machine vision system that is trained to output a probability of correct insertion based on the image subtraction results.In this or any other embodiment, the method further includes receiving cable insertion parameters prior to insertion of the cables, the cable insertion parameters specifying a proper insertion sequence for the one or more cable wires into the one or more cable cavities of the connector housing.

[0083] In an embodiment, a housing inspection system includes a controller configured to: receive an inspection image of a connector housing held in a housing holder of the housing inspection system from a camera system; classify the connector housing into an identified connector housing type via a connector recognition machine vision system based at least in part on a comparison between the inspection image of the connector housing and a template image corresponding to the identified connector housing type; and automatically verify, for one or more cable cavities of the connector housing, whether a correct cable is inserted into the cable cavity according to the identified connector housing type. In this or any other embodiment, the connector recognition machine vision system is trained with a plurality of different template images corresponding to a plurality of different connector housing types. In this or any other embodiment, the camera system is a stereo camera system including a first camera and a second camera, and wherein the connector recognition machine vision system includes a first recognition model trained to classify the inspection image output by the first camera and a second recognition model trained to classify the inspection image output by the second camera. In this or any other embodiment, the controller is further configured to detect the location of the one or more cable cavities in the inspection image by detecting a correspondence between image features in the inspection image and image features in the template image. In this or any other embodiment, the controller is further configured to generate a homography matrix to account for a rotation of the position of the one or more cable cavities caused by a rotation of the connector housing relative to the template image. In this or any other embodiment, a set of cables are inserted into the one or more cable cavities according to a series of one or more cable insertion steps of an insertion sequence, and wherein, after each insertion step, the controller is further configured to automatically verify that the most recently inserted cable is inserted into the correct one of the one or more cable cavities.

[0084] In an embodiment, a method for cable insertion verification includes: receiving an inspection image of a connector housing held in a shell holder of a shell inspection system from a camera system, wherein the camera system is attached to the shell inspection system at a fixed position relative to the shell holder; classifying the connector housing as an identified connector housing type via a connector recognition machine vision system by identifying a correspondence between a cable cavity detected in the inspection image and a cable cavity detected in the template image based at least in part on a comparison between the inspection image of the connector housing and a template image corresponding to the identified connector housing type; and automatically verifying whether a correct cable is inserted into the cable cavity of the connector housing according to the identified connector housing type after each insertion step in a series of insertion steps of an insertion sequence.

[0085] It will be understood that the configuration and / or manner described herein are exemplary in nature, and these specific embodiments or embodiments should not be considered to have a restrictive meaning because many changes can be made. The specific routine or method described herein can represent one or more of any number of processing strategies. Thus, the different actions shown and / or described can be performed or omitted in the order shown and / or described, in other orders, in parallel. Similarly, the order of the above-mentioned processing can be changed.

[0086] The subject matter of the present disclosure includes all novel and nonobvious combinations and subcombinations of the various processes, systems and configurations, and other features, functions, acts and / or properties of the present disclosure, as well as any and all equivalents thereof.

Claims

1. A method (200) for cable insertion verification, the method (200) comprising: receiving (202) an inspection image (404) of a connector housing (302) held in a housing holder (306) of a housing inspection system (300) from a camera system (402); classifying (204) the connector housing (302) as the identified connector housing type (410) based at least in part on a comparison between the inspection image (404) of the connector housing (302) and a template image (408) corresponding to the identified connector housing type (410), via a connector identification machine vision system (406); as well as For one or more cable cavities (304) of the connector housing (302), automatically verifying (206) whether a correct cable (108) is inserted into the cable cavity (304) based on the identified connector housing type (410).

2. The method (200) according to claim 1, wherein The connector recognition machine vision system (406) is trained with a plurality of different template images (408) corresponding to a plurality of different connector housing types.

3. The method (200) according to claim 2, wherein: The connector identification machine vision system (406) includes one of a support vector machine and an artificial neural network.

4. The method (200) according to claim 1, wherein The camera system (402) is a stereo camera system including a first camera (316A) and a second camera (316B), and wherein the connector identification machine vision system (406) includes a first recognition model and a second recognition model, wherein the first recognition model is trained to classify the inspection image output by the first camera (316A) and the second recognition model is trained to classify the inspection image output by the second camera (316B).

5. The method (200) according to claim 1, wherein The camera system (312) is attached to the housing inspection system (300) at a fixed position relative to the housing holder (306).

6. The method (200) according to claim 1, wherein The one or more cable cavities (304) of the connector housing (302) extend from an insertion face (320) of the connector housing (302) to a viewing face (322) of the connector housing (302), and wherein the inspection image (314) of the connector housing (302) depicts the viewing face (322).

7. The method (200) according to claim 6, wherein The housing inspection system (300) further includes an illumination system (318) configured to emit illumination light toward the viewing surface (322) of the connector housing (302).

8. The method (200) according to claim 1, further comprising: The positions of the one or more cable cavities (304) in the inspection image (314) are detected by detecting a correspondence between image features (504) in the inspection image (314) and image features (506) in the template image (502).

9. The method (200) according to claim 8, further comprising: A homography matrix (508) is generated to account for rotation of the position of the one or more cable cavities (304) caused by rotation of the connector housing (302) relative to the template image (502).

10. The method (200) according to claim 1, wherein A set of cables (108) are inserted into the one or more cable cavities (304) according to a series of one or more cable insertion steps (702) of an insertion sequence (700), and wherein, after each insertion step (702), the method further comprises automatically verifying whether the most recently inserted cable (108) is inserted into the correct cable cavity of the one or more cable cavities (304).

11. The method (200) according to claim 10, further comprising: A pre-insertion image (600) is captured before each insertion step (702), a post-insertion image (606) is captured after each insertion step (702), and an image subtraction result (612) is generated based on the pre-insertion image (600) and the post-insertion image (606).

12. The method (200) according to claim 11, wherein: The image subtraction result (612) is provided to an insertion verification machine vision system (614), which is trained to output a probability of correct insertion (616) based on the image subtraction result (612).

13. The method (200) of claim 1, further comprising, prior to the cable wires being inserted, receiving cable insertion parameters (310), the cable insertion parameters (310) specifying a correct insertion sequence for inserting one or more cable wires (108) into the one or more cable cavities (304) of the connector housing (302).

14. A shell inspection system (300), comprising: The controller (308) is configured to: receiving an inspection image (404) of a connector housing (302) held in a housing holder (306) of a housing inspection system (300) from a camera system (402); classifying, via a connector identification machine vision system (406), the connector housing (302) as the identified connector housing type (410) based at least in part on a comparison between the inspection image (404) of the connector housing (302) and a template image (408) corresponding to the identified connector housing type (410); as well as For one or more cable cavities (304) of the connector housing (302), automatically verifying whether a correct cable (108) is inserted into the cable cavity (304) based on the identified connector housing type (410).

15. The shell inspection system (300) according to claim 14, wherein: The connector recognition machine vision system (406) is trained with a plurality of different template images (408) corresponding to a plurality of different connector housing types.

16. The shell inspection system (300) according to claim 14, wherein: The camera system (402) is a stereo camera system including a first camera (316A) and a second camera (316B), and wherein the connector identification machine vision system (406) includes a first recognition model and a second recognition model, wherein the first recognition model is trained to classify the inspection image output by the first camera (316A) and the second recognition model is trained to classify the inspection image output by the second camera (316B).

17. The shell inspection system (300) according to claim 14, wherein: The controller (308) is further configured to detect the location of the one or more cable cavities (304) in the inspection image (404) by detecting a correspondence between image features (504) in the inspection image (404) and image features (506) in the template image (502).

18. The shell inspection system (300) according to claim 17, wherein: The controller (308) is further configured to generate a homography matrix (508) to account for a rotation of the position of the one or more cable cavities (304) caused by a rotation of the connector housing (302) relative to the template image (502).

19. The shell inspection system (300) according to claim 14, wherein: A set of cables (108) are inserted into the one or more cable cavities (304) according to a series of one or more cable insertion steps (702) of an insertion sequence (700), and wherein, after each insertion step (702), the controller (308) is further configured to automatically verify whether the most recently inserted cable (108) is inserted into the correct cable cavity of the one or more cable cavities (304).

20. A method (200) for cable insertion verification, the method (200) comprising: receiving (202) an inspection image (404) of a connector housing (302) held in a housing holder (306) of a housing inspection system (300) from a camera system (402), wherein the camera system (402) is attached to the housing inspection system (300) at a fixed position relative to the housing holder (306); classifying (204) the connector housing (302) as the identified connector housing type (410) by identifying a correspondence between a detected cable cavity (304) in the inspection image (404) and a cable cavity (304) in the template image (408), based at least in part on a comparison between the inspection image (404) of the connector housing (302) and a template image (408) corresponding to the identified connector housing type (410); and After each insertion step (702) in a series of insertion steps of an insertion sequence (700), it is automatically verified (206) whether the correct cable wire (108) is inserted into the cable cavity (304) of the connector housing (302) according to the identified connector housing type (410).

Images