A visual judgment wiring method and system for electrical assembly

Abstract

A visual judgment wiring method for electrical assembly, obtaining image information of a target box; dividing multiple ROI target areas according to the image information, and matching component templates from a template library with the image information of the ROI target areas; obtaining a box template pre-implanted with multiple preset calibration relationships of the component templates based on the component template; the box template adjusts the preset coordinate information of the component template located in the box template according to the image information of the ROI target area to generate the actual coordinate information of the device to be wired in the component template. The present invention detects the accuracy of wiring of large electric controls in real time, performs real-time coordinate correction on position-shifted devices, and adopts a modular independent matching correction method for components in the target box for actual position shift during the wiring process, so as to meet the requirements of automatic correction of template wiring points after detecting installation shift and process movement of the electric control, and greatly improves the qualified rate of one-time wiring of the electric control, thereby improving the wiring efficiency of employees.

Description

Technical Field

[0001] The invention relates to the technical field of assembly of electric control components of large air-conditioning units, and in particular to a visual judgment wiring method and system for electrical assembly. Background Art

[0002] The main problems in the assembly process of electronic control components of large commercial air-conditioning units are as follows: the wiring guidance and wiring inspection process are traditional, and the wiring is guided manually through paper circuit diagrams throughout the process, and the wiring accuracy is determined by manual inspection and marking; the detection rate of wiring abnormalities is low, there is no data record of the wiring process, and quality traceability is impossible. The minimum wire diameter of all electronic control wiring is 0.5mm, the line number characters are small, and the efficiency of inputting wiring information one by one is low. The photos taken by the industrial camera at a set distance have blurred silk screen printing on the components of the electronic control cabinet and there are different degrees of position offset changes during installation. The product position is offset and it is difficult to re-correct; the employee skills requirements are high, the training cycle is long, the electrical expertise must be strong, and the quality requirements of the operation process are high; the efficiency of manual reading and determination of wiring position is low, and there are more than 450 wires and more than 900 wiring points. It takes 80 to 90 minutes to check the dual system and 40 to 50 minutes for the single system.

[0003] Therefore, there are the following technical difficulties in the electrical assembly industry of large commercial air-conditioning units:

[0004] 1. There are at least 100 wirings for a single electronic control component. Inputting wiring information into the system one by one is extremely inefficient and prone to errors.

[0005] 2. The photos taken by the industrial camera (20 million) at the corresponding set distance have blurred silk screens on the components of the electric control cabinet, which is difficult to see clearly, reducing the efficiency of drawing the wiring position area during the modeling process. The wiring positions are staggered and the plane is blocked and cannot be judged;

[0006] 3. The actual position of the box deviates from the template position, and the template cannot be matched during the detection process. Some components such as relays and circuit breakers are installed on rails, and each installation has different degrees of position offset changes. The template cannot be matched during the system detection process. Once the product position is offset, it is difficult to re-correct it.

[0007] Therefore, a new process method for guiding and testing the electrical wiring of large commercial air-conditioning units was developed. A local feature recognition algorithm was used to achieve one-time matching of the external frame of the box with the template before wiring, and real-time matching of the devices inside the box. This improves the wiring efficiency of employees, optimizes the wiring error-proofing process, and increases quality assurance in the electrical control assembly production stage. At the same time, the system has the function of querying detection records to improve the technological advancement. Summary of the invention

[0008] In response to one or more of the above-mentioned problems of lack of accuracy in wiring judgment, low efficiency, low detection rate of wiring anomalies, etc., the present invention provides a visual wiring judgment method and system for electrical assembly, which adopts local feature recognition to achieve one-time matching of the target box outer frame with the box template before wiring, and real-time matching of components or devices in the target box, thereby improving the wiring efficiency of employees.

[0009] To achieve the above object, the present invention adopts the following technical solution: a visual judgment wiring method for electrical assembly, comprising:

[0010] Get the image information of the target box;

[0011] Divide a plurality of ROI target areas according to the image information, and match component templates from a template library with the image information of the ROI target areas;

[0012] Based on the component template, a box template pre-implanted with a plurality of preset calibration relationships of the component templates is obtained, wherein the plurality of preset calibration relationships of the component templates at least include an identifier of each component template and a position mapping relationship between the identifier and the box template;

[0013] The box template adjusts the preset coordinate information of the component template located in the box template according to the image information of the ROI target area to generate actual coordinate information of the device to be wired in the component template.

[0014] Preferably, the visual judgment wiring method further comprises:

[0015] Acquire image information of the target box, including acquiring image information around the outside of the target box;

[0016] Determine and determine the target box offset based on the preset coordinate information of the box template and the image information around the target box body;

[0017] The preset coordinate information of the box template is adjusted according to the target box offset, the actual coordinate information of the box template is obtained, and the positioning of the box template is corrected so that the box template matches the target box.

[0018] Preferably, the step of obtaining a box template pre-implanted with a plurality of preset calibration relationships of the component templates comprises:

[0019] Matching corresponding component templates from a template library according to the image information of the multiple ROI target areas, obtaining a unique identifier of each component template, and determining a position mapping relationship between each component template and / or between each component template and a box template;

[0020] The preset calibration relationship is determined by the unique identification and position mapping relationship of each component template, and the cabinet template is matched from the template library according to the preset calibration relationship, wherein the position mapping relationship is set to allow the component offset when matching the cabinet template to improve the adaptability when matching the cabinet template;

[0021] The box template is pre-implanted in the template library, and when implanted, a preset calibration relationship with the selection combination of various component templates and the position layout of the component template is created, so that when matching, the box template with the corresponding preset calibration relationship is determined based on the unique identification and position mapping relationship of each component template.

[0022] Preferably, dividing a plurality of ROI target areas according to the image information comprises:

[0023] Identify the position of the end of the device to be wired according to the image information of the target box;

[0024] Extending outwards from the end position of the device to be wired as the center to form a dividing boundary, wherein the dividing boundary covers a target area of ​​at least one device to be wired;

[0025] The target region defined by the division boundary is extracted to form a ROI target region.

[0026] Preferably, the actual coordinate information of the components to be wired in the generated component template includes:

[0027] According to the image information of the ROI target area, specific position information of each ROI target area within the target box area is obtained;

[0028] The target component offset is determined by judging the specific position information of the ROI target area and the preset coordinate information of the component template;

[0029] Adjusting the preset coordinate information of the component template in the box template by the target component offset, obtaining the actual coordinate information of the component template and correcting the positioning of the component template;

[0030] The actual coordinate information of the component template is used as a reference to obtain the actual coordinate information of the device to be wired in the component template.

[0031] Preferably, judging and determining the target component offset by using the specific position information of the ROI target area and the preset coordinate information of the component template comprises:

[0032] The ROI target area matches the component template from the template library with the device end to be wired, and loads the component template;

[0033] Loading the position of the component template located in the box template and the specific position information of the ROI target area to form a target component offset, wherein the target component offset includes a displacement in the front, back, left, and right directions based on the preset coordinate information of the component template located in the box template and an angular offset centered on the end of the device to be wired;

[0034] The component templates are individually adjusted according to the target offset to correct the positioning.

[0035] Preferably, the visual judgment wiring method further comprises:

[0036] Acquire image information of the target box, wherein the target box is a target box with a mainboard structure of more than two layers;

[0037] Obtain the lower layer image information of the target box and divide the ROI target area;

[0038] Matching the component template and the box template according to the ROI target area and adjusting according to the target component offset and / or the target box offset to correct the positioning;

[0039] After the lower layer is assembled, the image information of the upper layer adjacent to the lower layer is obtained and the positioning is corrected until the multi-layer structure assembly of the target box is completed.

[0040] Preferably, the step of acquiring image information of the target box includes:

[0041] Arrange binocular vision recognition units according to the target box, and obtain first image information and second image information of the target box through one vision recognition unit and another vision recognition unit respectively;

[0042] Performing barrel transformation on the first image information and / or the second image information acquired by a visual recognition unit to reduce the divergence of edge image matching and the deformation after perspective transformation so as to reduce the distortion of the image information after splicing;

[0043] Performing feature point matching on the first image information and / or the second image information acquired by the binocular vision recognition unit, and automatically identifying matching points using a perspective matrix;

[0044] The first image information and / or the second image information acquired by another visual recognition unit is perspective transformed and spliced ​​with another image information to obtain image information of the target box.

[0045] On the other hand, the present invention also provides the following technical solution: a visual judgment wiring system for electrical assembly, comprising:

[0046] An image acquisition unit, used to acquire image information of the surrounding and internal areas of the target box;

[0047] A region division unit, used for dividing a plurality of ROI target regions according to the positions of the ends of the devices to be connected according to the image information;

[0048] A template matching unit, used for matching component templates from a database according to the plurality of ROI target areas and matching a box template according to a preset calibration relationship of the component template;

[0049] A database for storing pre-implanted component templates and cabinet templates;

[0050] The correction positioning unit is used to calculate the offset of the component template and the box template according to the image information of the ROI target area, and correct the positioning of the component template and the box template according to the offset.

[0051] Preferably, the image acquisition unit comprises:

[0052] A binocular vision recognition unit, having a left, right or top, bottom arranged vision recognition unit, for acquiring first image information and second image information of a target box;

[0053] An image information transformation unit, used to perform barrel transformation on the first image information and / or the second image information acquired by a visual recognition unit, so as to reduce the divergence of edge image matching and the deformation after perspective transformation so as to reduce the distortion of the image information after splicing;

[0054] A feature point matching unit, used to perform feature point matching on the first image information and / or the second image information acquired by the binocular vision recognition unit, and automatically identify matching points using a perspective matrix;

[0055] The image information splicing unit is used to perform perspective transformation on the first image information and / or the second image information acquired by another visual recognition unit, and splice them with another image information to obtain image information of the target box.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] In order to meet the depth of field requirements, the present invention uses a binocular vision recognition unit to collect the center of the target box area, takes the same image as the matching center line, and then restores the image position information according to the binocular vision positioning mathematical model after segmenting the first image and the second image, completes the real-time detection and positioning of the target box, realizes arc high-precision line number recognition, and meets 0.6mm 2 Accurately identify the line number on the outer diameter line, use the image domain segmentation data processing algorithm, and perform real-time coordinate correction on the position offset device to meet the needs of automatic correction of the template wiring points after the installation offset and process movement of the detection electronic control.

[0058] In addition to establishing a placement reference for the target box, the present invention also performs adaptive calculations in the algorithm for the actual position offset during the wiring process. It abandons the overall structure secondary matching method with a huge amount of calculation, and adopts a modular independent matching correction method for components in the target box. The ROI target area is obtained through the line end position of the device to be connected, and the local device correction positioning is achieved through component search. The actual coordinate information of the device to be connected in the component template is used as a reference to obtain the actual coordinate information of the device to be connected in the component template.

[0059] The present invention adopts local feature recognition, detects multiple ROI target areas during data matching, and allows relative movement (displacement and rotation) between matching devices, that is, the target component offset, and adjusts the preset coordinate information of the component template located in the box template with the target component offset, obtains the actual coordinate information of the component template and corrects the positioning of the component template, solves the difficulty of correcting the local offset of the guide rail mounted device, and at the same time solves the problem of affecting the coordinate parameters of other fixed devices during correction, detects the accuracy of large electric control wiring in real time, and completely eliminates the tedious steps of double marking (self-inspection, mutual inspection) inspection of traditional connectors, and the qualified rate of one-time wiring of electric control is increased to more than 99%, without the phenomenon of wrong connection judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solution, the drawings required for use in the implementation will be briefly introduced below. Obviously, the drawings described below are only some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0061] Figure 1 It is a structural schematic diagram of the present invention. DETAILED DESCRIPTION

[0062] In order to enable a clear and complete understanding of the technical solution, the present invention is further described in conjunction with the embodiments and drawings. Obviously, the described embodiments are only some embodiments of the present invention, and all other embodiments obtained by technicians in the relevant field without making creative work are within the scope of protection of the present invention.

[0063] like Figure 1 As shown, a method for visually judging wiring of electrical assembly includes:

[0064] S100, acquiring image information of a target box;

[0065] S110, arranging a binocular vision recognition unit according to the target box, including a left camera and a right camera, and taking pictures of the target box by the left camera and the right camera respectively to obtain first image information and second image information respectively;

[0066] S120, performing barrel transformation on the first image information and / or the second image to reduce the divergence of edge image matching and the deformation after perspective transformation, thereby reducing the distortion of the image information after stitching;

[0067] S130, performing feature point matching on the first image information and / or the second image information acquired by the left camera and the right camera, and automatically identifying matching points using a perspective matrix;

[0068] S140, performing perspective transformation on the first image information and / or the second image information and then splicing them to achieve image fusion to obtain image information of the target box.

[0069] In this implementation case, the lens concentration is selected to be 28mm, the object distance is 1650mm, and the depth of field range is 280mm to meet the depth of field requirements. The binocular vision recognition unit is used to collect the center of the target box area, and the same image is used as the matching center line. After the first image and the second image are segmented, the image position information is restored according to the binocular vision positioning mathematical model to complete the real-time detection and positioning of the target box, realize the arc high-precision line number recognition, and meet the 0.6mm depth requirement. 2 The line number above the outer diameter line can be accurately identified.

[0070] At present, the target cabinets (large electric control cabinets) involve hundreds of specifications and types. The camera field of view needs to exceed 2700mm (length) × 1600mm (width) to cover all products. At the same time, the large electric control cabinet has large differences in the size and specifications of internal components, with a height difference of nearly 200mm. The minimum cross-sectional area of ​​the internal copper core of the wiring is 0.5mm 2 When the camera takes photos, its depth of field and pixel accuracy must be guaranteed to meet the image recognition of all wiring diameters, so a normal single 20-megapixel industrial camera cannot meet the detection requirements. This implementation case uses two 20-megapixel cameras or cameras to complete the global coverage of the large electric control cabinet, and then uses image stitching technology to achieve image fusion to obtain the image information of the target cabinet.

[0071] S200, dividing a plurality of ROI target regions according to the image information, and matching component templates from a template library according to the image information of the ROI target regions;

[0072] S210, identifying the position of the end of the device to be wired according to the image information of the target box;

[0073] S220, extending outwards with the end position of the device to be wired as the center to form a dividing boundary, wherein the dividing boundary covers a target area of ​​at least one device to be wired;

[0074] S230: extracting a target region defined by the division boundary to form a ROI target region.

[0075] S240, matching corresponding component templates from a template library according to the image information of the plurality of ROI target areas, obtaining a unique identifier of each component template, and determining a position mapping relationship between each component template and / or between each component template and a box template;

[0076] S250, in order to improve the efficiency of matching component templates and reduce the amount of calculation, the allowable component offset is set during matching. The allowable component offset allows the component to be offset by 30mm in the left and right direction, 20mm in the front and back direction, and an angle offset of ±20°. Local device correction positioning is achieved during component search and matching, and the amount of calculation is reduced by 80% compared to overall matching.

[0077] S300, based on the component template, obtaining a box template pre-implanted with a plurality of preset calibration relationships of the component templates, wherein the plurality of preset calibration relationships of the component templates at least include an identifier of each component template and a position mapping relationship between the identifier and the box template;

[0078] S310, determining a preset calibration relationship based on the unique identification and position mapping relationship of each component template, that is, selecting and combining the types, models and wiring methods of each component template, and the position layout between the component templates and between the box templates, to obtain a preset calibration relationship of the combination and position relationship of multiple component templates;

[0079] S320, matching a cabinet template from a template library according to a preset calibration relationship, wherein a position mapping relationship is set when matching the cabinet template to allow a component offset to improve the adaptability when matching the cabinet template; the cabinet template is pre-implanted in the template library, and a preset calibration relationship with a selection combination of various component templates and a position layout of the component template is created when implanted, so that a cabinet template with a corresponding preset calibration relationship is determined based on the unique identifier and position mapping relationship of each component template during matching;

[0080] S330, after obtaining the box template, obtain the image information of the outer periphery of the target box through the image information of the target box, and judge and determine the target box offset according to the preset coordinate information of the box template and the image information of the outer periphery of the target box. The target box offset allows an offset of 200 mm in the front, back, left, and right directions and an angle offset of ±40°;

[0081] S340, adjusting the preset coordinate information of the box template with the target box offset, obtaining the actual coordinate information of the box template and correcting the positioning of the box template so that the box template matches the target box.

[0082] S400, the box template adjusts the preset coordinate information of the component template located in the box template according to the image information of the ROI target area to generate actual coordinate information of the device to be wired in the component template;

[0083] S410, obtaining specific position information of each ROI target area within the target box area according to the image information of the ROI target area;

[0084] S420, judging and determining the target component offset according to the specific position information of the ROI target area and the preset coordinate information of the component template;

[0085] S421, the ROI target area matches a component template from a template library with the device end to be wired, and loads the component template, and the position of the loaded component template located in the box template and the specific position information of the ROI target area form a target component offset;

[0086] S422, the target component offset includes the displacement in the front, back, left, and right directions based on the preset coordinate information of the component template located on the box template and the angular offset centered on the end of the device to be wired, wherein the target component offset allows an offset of 30 mm in the left and right direction, 20 mm in the front and back direction, and an angular offset of ±20°;

[0087] S423, adjusting the component template separately according to the target offset to correct the positioning. In this implementation case, in addition to establishing the placement reference of the target box, the algorithm also performs adaptive calculations for the actual position offset during the wiring process, abandoning the overall structure secondary matching method with huge computational complexity, and adopting a modular independent matching correction method for the components in the target box. The ROI target area is obtained through the line end position of the device to be connected, and the local device correction positioning is achieved through component search. The actual coordinate information of the device to be connected in the component template is used as a reference to obtain the actual coordinate information of the device to be connected in the component template.

[0088] This implementation case is to solve the problem that the template cannot be matched during the traditional system detection process when the device is offset. There are two main types of device offsets:

[0089] 1. Some components such as relays and circuit breakers are installed on rails. Each installation will cause position deviation to varying degrees, and the system detection process template cannot match;

[0090] 2. The actual position of the box deviates and shifts in real time, and the system judgment process will cause wiring misjudgment.

[0091] In response to the above problems, this implementation case adopts a component template matching algorithm in the overall matching system of the large electric control cabinet. During data matching, multiple ROI target areas are detected and relative movement (displacement and rotation) between matching devices is allowed, that is, the target component offset. The preset coordinate information of the component template located in the box template is adjusted by the target component offset to obtain the actual coordinate information of the component template and correct the positioning of the component template, thereby solving the difficulty in correcting the local offset of the guide rail mounted device and solving the problem of affecting the coordinate parameters of other fixed devices during correction.

[0092] As a further improvement of the present implementation case: the target box has a double-layer mainboard structure design inside, and the double-layer mainboard structure identification and detection, modeling and wiring guidance cannot be implemented in the same template. This structure adopts the upper and lower layer sub-template method during modeling, and the identification method is to identify the lower layer first, and then start the upper layer assembly after the lower layer assembly is completed. Specifically, it includes obtaining the lower layer image information of the target box and dividing the ROI target area, matching the component template and the box template according to the ROI target area, and adjusting the positioning according to the target component offset and / or the target box offset, and after the lower layer assembly is completed, obtaining the upper layer image information adjacent to the lower layer image information and correcting the positioning, until the multi-layer structure assembly of the target box is completed, to ensure the feasibility of the electronic control production operation and the product quality.

[0093] As a further improvement of this implementation case: some of the ends of the devices to be connected are common-voltage terminal structures, and the common-voltage terminal identification and detection cannot guarantee the wiring order of the terminals. This implementation case adopts the method of identifying one of the line numbers, and the system establishes a multi-line sequential wiring guidance mode. The system automatically brings out all the wiring information, and sorts the wiring for all the wiring, connects and confirms in sequence, and realizes wiring guidance and accuracy judgment.

[0094] The present invention also provides another implementation example: a visual judgment wiring system for electrical assembly, comprising:

[0095] An image acquisition unit is used to acquire image information of the surrounding and internal areas of the target box, and the image acquisition unit includes:

[0096] A binocular vision recognition unit, having a left, right or top, bottom arranged vision recognition unit, for acquiring first image information and second image information of a target box;

[0097] An image information transformation unit, used to perform barrel transformation on the first image information and / or the second image information acquired by a visual recognition unit, so as to reduce the divergence of edge image matching and the deformation after perspective transformation so as to reduce the distortion of the image information after splicing;

[0098] A feature point matching unit, used to perform feature point matching on the first image information and / or the second image information acquired by the binocular vision recognition unit, and automatically identify matching points using a perspective matrix;

[0099] An image information splicing unit, used to perform perspective transformation on the first image information and / or the second image information acquired by another visual recognition unit, and splice the first image information and / or the second image information with another image information to obtain image information of the target box;

[0100] A region division unit, used for dividing a plurality of ROI target regions according to the positions of the ends of the devices to be connected according to the image information;

[0101] A template matching unit, used for matching component templates from a database according to the plurality of ROI target areas and matching a box template according to a preset calibration relationship of the component template;

[0102] A database for storing pre-implanted component templates and cabinet templates;

[0103] The correction positioning unit is used to calculate the offset of the component template and the box template according to the image information of the ROI target area, and correct the positioning of the component template and the box template according to the offset.

[0104] The above disclosure is only one or more preferred embodiments of the present invention, which are used to help understand the inventive concept of the technical solution, and is not intended to limit the present invention in other forms. Technical personnel in the relevant field may make other equivalent or customary replacement schemes based on the features defined by the present invention, which still fall within the scope of the present invention.

Claims

1. A visual judgment wiring method for electrical assembly, characterized in that include: Get the image information of the target box; Divide a plurality of ROI target areas according to the image information, and match component templates from a template library with the image information of the ROI target areas; Based on the component template, a box template pre-implanted with a plurality of preset calibration relationships of the component templates is obtained, wherein the plurality of preset calibration relationships of the component templates at least include an identifier of each component template and a position mapping relationship between the identifier and the box template; The box template adjusts the preset coordinate information of the component template located in the box template according to the image information of the ROI target area to generate actual coordinate information of the device to be wired in the component template.

2. A visual judgment wiring method for electrical assembly according to claim 1, characterized in that: The visual judgment wiring method also includes: Acquire image information of the target box, including acquiring image information around the outside of the target box; Determine and determine the target box offset based on the preset coordinate information of the box template and the image information around the target box body; The preset coordinate information of the box template is adjusted according to the target box offset, the actual coordinate information of the box template is obtained, and the positioning of the box template is corrected so that the box template matches the target box.

3. A visual judgment wiring method for electrical assembly according to claim 1, characterized in that: The step of obtaining a box template pre-implanted with a plurality of component templates with preset calibration relationships includes: Matching corresponding component templates from a template library according to the image information of the multiple ROI target areas, obtaining a unique identifier of each component template, and determining a position mapping relationship between each component template and / or between each component template and a box template; The preset calibration relationship is determined by the unique identification and position mapping relationship of each component template, and the cabinet template is matched from the template library according to the preset calibration relationship, wherein the position mapping relationship is set to allow the component offset when matching the cabinet template to improve the adaptability when matching the cabinet template; The box template is pre-implanted in the template library, and when implanted, a preset calibration relationship with the selection combination of various component templates and the position layout of the component template is created, so that when matching, the box template with the corresponding preset calibration relationship is determined based on the unique identification and position mapping relationship of each component template.

4. A method for visually judging wiring of electrical assembly according to claim 1, characterized in that: The dividing of a plurality of ROI target areas according to the image information comprises: Identify the position of the end of the device to be wired according to the image information of the target box; Extending outwards from the end position of the device to be wired as the center to form a dividing boundary, wherein the dividing boundary covers a target area of ​​at least one device to be wired; The target region defined by the division boundary is extracted to form a ROI target region.

5. A visual judgment wiring method for electrical assembly according to claim 4, characterized in that: The actual coordinate information of the components to be wired in the generated component template includes: According to the image information of the ROI target area, specific position information of each ROI target area within the target box area is obtained; The target component offset is determined by judging the specific position information of the ROI target area and the preset coordinate information of the component template; Adjusting the preset coordinate information of the component template in the box template by the target component offset, obtaining the actual coordinate information of the component template and correcting the positioning of the component template; The actual coordinate information of the component template is used as a reference to obtain the actual coordinate information of the device to be wired in the component template.

6. A method for visually judging wiring of electrical assembly according to claim 5, characterized in that: The step of judging and determining the target component offset by using the specific position information of the ROI target area and the preset coordinate information of the component template includes: The ROI target area matches the component template from the template library with the device end to be wired, and loads the component template; Loading the position of the component template located in the box template and the specific position information of the ROI target area to form a target component offset, wherein the target component offset includes a displacement in the front, back, left, and right directions based on the preset coordinate information of the component template located in the box template and an angular offset centered on the end of the device to be wired; The component templates are individually adjusted according to the target offset to correct the positioning.

7. A method for visually judging wiring of electrical assembly according to claim 1, characterized in that: The visual judgment wiring method also includes: Acquire image information of the target box, wherein the target box is a target box with a mainboard structure of more than two layers; Obtain the lower layer image information of the target box and divide the ROI target area; Matching the component template and the box template according to the ROI target area and adjusting according to the target component offset and / or the target box offset to correct the positioning; After the lower layer is assembled, the image information of the upper layer adjacent to the lower layer is obtained and the positioning is corrected until the multi-layer structure assembly of the target box is completed.

8. A method for visually judging wiring of electrical assembly according to any one of claims 1 to 7, characterized in that: The obtaining of the image information of the target box includes: Arrange binocular vision recognition units according to the target box, and obtain first image information and second image information of the target box through one vision recognition unit and another vision recognition unit respectively; Performing barrel transformation on the first image information and / or the second image information acquired by a visual recognition unit to reduce the divergence of edge image matching and the deformation after perspective transformation so as to reduce the distortion of the image information after splicing; Performing feature point matching on the first image information and / or the second image information acquired by the binocular vision recognition unit, and automatically identifying matching points using a perspective matrix; The first image information and / or the second image information acquired by another visual recognition unit is perspective transformed and spliced ​​with another image information to obtain image information of the target box.

9. A visual judgment wiring system for electrical assembly, characterized in that include: An image acquisition unit, used to acquire image information of the surrounding and internal areas of the target box; A region division unit, used for dividing a plurality of ROI target regions according to the positions of the ends of the devices to be connected according to the image information; A template matching unit, used for matching component templates from a database according to the plurality of ROI target areas and matching a cabinet template according to a preset calibration relationship of the component template; A database for storing pre-implanted component templates and cabinet templates; The correction positioning unit is used to calculate the offset of the component template and the box template according to the image information of the ROI target area, and correct the positioning of the component template and the box template according to the offset.

10. A visual judgment wiring system for electrical assembly according to claim 9, characterized in that: The image acquisition unit comprises: A binocular vision recognition unit, having a left, right or top, bottom arranged vision recognition unit, for acquiring first image information and second image information of a target box; An image information transformation unit, used to perform barrel transformation on the first image information and / or the second image information acquired by a visual recognition unit, so as to reduce the divergence of edge image matching and the deformation after perspective transformation so as to reduce the distortion of the image information after splicing; A feature point matching unit, used to perform feature point matching on the first image information and / or the second image information acquired by the binocular vision recognition unit, and automatically identify matching points using a perspective matrix; The image information splicing unit is used to perform perspective transformation on the first image information and / or the second image information acquired by another visual recognition unit, and splice them with another image information to obtain image information of the target box.

Images