Automobile part identification and positioning method and system based on machine vision

Abstract

The invention provides an automobile part recognition and positioning method and system based on machine vision, and the method comprises the steps: carrying out the multi-angle image collection of an automobile part based on a preset camera, obtaining a multi-angle part image, and carrying out the standardization processing of the multi-angle part image; preprocessing the standardized multi-angle component images, and respectively extracting component characterization features in the multi-angle component images based on a preprocessing result; and identifying and positioning the automobile parts based on the part characterization features, and obtaining specific position coordinates of the automobile parts based on an identifying and positioning result. Effective recognition and positioning of the automobile parts are achieved, the accuracy and effectiveness of determining the specific position coordinates of the automobile parts are improved, and the accuracy and efficiency of recognition and positioning of the automobile parts are guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a method and system for identifying and locating automobile parts based on machine vision. Background Art

[0002] At present, with the rapid development of the automotive industry, the efficiency and precision requirements for the production and assembly of automotive parts are becoming increasingly higher. The continuous advancement of machine vision technology has provided a new way to solve these problems.

[0003] However, traditional auto parts identification and positioning mostly use a combination of manual and machine recognition. Due to the use of manual recognition, the identification and positioning of parts are inefficient, prone to errors, and costly, which greatly reduces the effectiveness of auto parts identification and positioning.

[0004] Therefore, in order to overcome the above-mentioned defects, the present invention provides a method and system for identifying and locating automobile parts based on machine vision. Summary of the Invention

[0005] The present invention provides a method and system for identifying and locating automobile parts based on machine vision, which is used to collect multi-angle images of automobile parts through a preset camera, and standardize the collected multi-angle component images, so as to facilitate effective preprocessing of the multi-angle component images. Secondly, component representation features are extracted from the preprocessed multi-angle component images, so as to facilitate identification and positioning of automobile parts based on the extracted component representation features. Finally, by analyzing the component representation features, effective identification and positioning of automobile parts are achieved, thereby improving the accuracy and effectiveness of determining the specific position coordinates of automobile parts, and ensuring the accuracy and efficiency of identification and positioning of automobile parts.

[0006] The present invention provides a method for identifying and locating automobile parts based on machine vision, comprising:

[0007] Step 1: Capture multi-angle images of auto parts using a preset camera to obtain multi-angle component images, and perform standardization on the multi-angle component images;

[0008] Step 2: Preprocess the standardized multi-angle component images, and extract component representation features in the multi-angle component images based on the preprocessing results;

[0009] Step 3: Identify and locate the automobile parts based on the component characterization features, and obtain the specific position coordinates of the automobile parts based on the identification and positioning results.

[0010] Preferably, a method for identifying and locating automobile parts based on machine vision, in step 1, before collecting multi-angle images of automobile parts based on a preset camera, includes:

[0011] Obtain image acquisition requirements for automotive parts, and based on the image acquisition requirements, determine acquisition angles for different types of automotive parts and preset camera acquisition postures at each acquisition angle;

[0012] Record the acquisition angles of different types of automotive parts and the corresponding acquisition postures of preset cameras, and generate a monitoring layout guidance strategy library for preset cameras based on the recording results;

[0013] Generate category identifiers based on category information of automobile parts, and determine target addresses of monitoring layout guidelines for different categories of automobile parts in a monitoring layout guide strategy library;

[0014] Based on the target address, a jump link is configured for different category identifiers, and based on the jump link, the access configuration of the category identifier and the corresponding monitoring layout guide in the monitoring layout guide policy library is completed.

[0015] Preferably, a method for identifying and locating automobile parts based on machine vision, in step 1, multi-angle image acquisition of automobile parts based on a preset camera includes:

[0016] Obtain the acquisition angles of automotive parts and determine the execution order between different acquisition angles based on acquisition requirements;

[0017] At the same time, based on the acquisition requirements, the adjustable range of the preset camera at each acquisition angle is determined, and the target shooting frequency of the preset camera at each acquisition posture within the adjustable range is determined;

[0018] Based on the target shooting frequency, the preset camera is controlled to capture images of automobile parts at corresponding acquisition angles according to the execution order to obtain multi-angle component images.

[0019] Preferably, a method for identifying and locating automobile parts based on machine vision, which obtains multi-angle component images, includes:

[0020] Acquire the obtained multi-angle component images and extract the shooting sources of the multi-angle component images;

[0021] Generate location tags for component images at different angles based on the shooting source, and perform location marking on the corresponding component images at different angles based on the location tags;

[0022] The multi-angle component images after position marking are classified by angle, and based on the classification results, the transmission queues of the automobile component images acquired at different angles are allocated;

[0023] Based on the allocation results, the images of automobile parts under different acquisition angles are transmitted to the management terminal in turn.

[0024] Preferably, a method for identifying and locating automobile parts based on machine vision, in step 1, standardizing multi-angle component images, includes:

[0025] Obtain the imaging coordinate system of the multi-angle component image at the corresponding acquisition angle, and at the same time, determine the target unified coordinate system for the multi-angle component image;

[0026] Determine the relative relationship between the imaging coordinate system at different acquisition angles and the target unified coordinate system, and determine the difference change scheme of the imaging coordinate system at different acquisition angles based on the relative relationship;

[0027] Based on the difference change scheme, the imaging coordinate systems under different acquisition angles are transformed, and a unified component image under the target unified coordinate system is obtained based on the coordinate transformation results;

[0028] The unified component image is sampled, and the sampling result is quality evaluated based on the image quality evaluation index. When the quality evaluation result does not meet the preset requirements, the unified component image is adjusted in multiple dimensions until the preset requirements are met.

[0029] Preferably, in a method for identifying and locating automobile parts based on machine vision, in step 2, preprocessing the standardized multi-angle component images comprises:

[0030] Read the standardized multi-angle component images, and perform multi-dimensional image quality assessment on each angle component image according to the image quality assessment dimension;

[0031] Output the quality assessment result of each angular component image according to the assessment result, read the quality assessment result, and determine the assessment value corresponding to each assessment dimension in each angular component image;

[0032] Sort the evaluation values ​​corresponding to each evaluation dimension in each angle component image in ascending order, and generate a dimension processing priority for each angle component image according to the sorting result;

[0033] Obtaining the dimension weight of the quality assessment dimension, and adapting the dimension weight of the quality assessment dimension to the dimension processing priority corresponding to each angular component image, respectively, to obtain the degree of adaptation between the dimension weight of the quality assessment dimension and the dimension processing priority corresponding to each angular component image;

[0034] Sorting the degree of fit between the dimension weights of the quality assessment dimensions and the dimension processing priorities corresponding to each angular component image from large to small, and obtaining the image processing priorities of the plurality of angular component images according to the sorting results;

[0035] Generate a first preprocessing strategy based on the image processing priorities of the plurality of angular component images, and generate a second preprocessing strategy based on the dimensional processing priority corresponding to each angular component image;

[0036] The second preprocessing strategy is integrated into the first preprocessing strategy to generate a comprehensive preprocessing strategy for preprocessing the standardized multi-angle component image;

[0037] The standardized multi-angle component images are preprocessed based on a comprehensive preprocessing strategy.

[0038] Preferably, a method for identifying and locating automobile parts based on machine vision, in step 2, extracting component representation features from multi-angle component images based on preprocessing results, includes:

[0039] Acquire the pre-processed multi-angle component image, analyze the multi-angle component image, and determine the depth information of the multi-angle component image;

[0040] Determine a target distance of the automobile part relative to a preset camera based on the depth information, determine an imaging size of the automobile part based on the target distance, and lock the imaging size;

[0041] Based on the locking results, pixel points of the multi-angle component images under corresponding imaging sizes are traversed respectively, and local features in the multi-angle component images are determined respectively based on the traversal results;

[0042] Obtain corresponding feature points based on local features, and perform regional segmentation on multi-angle component images based on the feature points;

[0043] Extract sensitive image areas of auto parts based on the region segmentation results, perform edge detection on the sensitive image areas, and determine the contour features of auto parts;

[0044] At the same time, pixel analysis is performed on sensitive image areas to determine the phenomenon representation features of automobile parts, and the contour features of automobile parts are summarized with the phenomenon representation features of automobile parts to obtain the component representation features in multi-angle component images.

[0045] Preferably, a method for identifying and locating automobile parts based on machine vision, in step 3, identifying and locating automobile parts based on the part characterization features, and obtaining the specific position coordinates of the automobile parts based on the identification and positioning results, includes:

[0046] Obtaining the obtained component characterization features, matching the component characterization features with preset component features, and determining the part category of the automotive component based on the matching results;

[0047] Extract the functional attributes of part categories and determine the distribution range of automotive parts based on the functional attributes;

[0048] Retrieving a local distribution map of automotive components based on the distribution range, and comparing the component characterization features with the known components in the local distribution map of automotive components for similarity;

[0049] Based on the similarity comparison result, the target position of the automobile parts in the local distribution map of automobile components is determined, and the coordinate transformation of the target position is performed based on the position relationship between the local distribution map of automobile components and the three-dimensional automobile coordinate position to obtain the specific position coordinates of the automobile parts.

[0050] The present invention provides an automobile parts identification and positioning system based on machine vision, comprising:

[0051] Image acquisition and standardization module, used to acquire multi-angle images of automotive parts based on preset cameras, obtain multi-angle component images, and perform standardization on the multi-angle component images;

[0052] A feature extraction module is used to preprocess the standardized multi-angle component images and extract component representation features in the multi-angle component images based on the preprocessing results;

[0053] The identification and positioning module is used to identify and locate automobile parts based on the component characterization features, and obtain the specific position coordinates of the automobile parts based on the identification and positioning results.

[0054] Preferably, a machine vision-based automobile parts identification and positioning system, image acquisition and standardization module, includes:

[0055] A configuration unit is used to obtain image acquisition requirements for automobile parts, and based on the image acquisition requirements, determine acquisition angles for different types of automobile parts and preset camera acquisition postures at each acquisition angle;

[0056] A monitoring layout guidance construction unit is used to record the acquisition angles of different types of automotive parts and the corresponding acquisition postures of preset cameras, and generate a monitoring layout guidance strategy library for preset cameras based on the recording results;

[0057] Monitoring layout guidance reference units, including:

[0058] Generate category identifiers based on category information of automobile parts, and determine target addresses of monitoring layout guidelines for different categories of automobile parts in a monitoring layout guide strategy library;

[0059] Based on the target address, a jump link is configured for different category identifiers, and based on the jump link, the access configuration of the category identifier and the corresponding monitoring layout guide in the monitoring layout guide policy library is completed.

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] 1. Multi-angle images of auto parts are captured by a preset camera, and the captured multi-angle component images are standardized, thereby facilitating effective preprocessing of the multi-angle component images. Secondly, component representation features are extracted from the preprocessed multi-angle component images, thereby facilitating identification and positioning of auto parts based on the extracted component representation features. Finally, by analyzing the component representation features, effective identification and positioning of auto parts are achieved, thereby improving the accuracy and effectiveness of determining the specific position coordinates of auto parts, and ensuring the accuracy and efficiency of identification and positioning of auto parts.

[0062] 2. By analyzing the image acquisition requirements of automobile parts, the acquisition angles of different categories of automobile parts and the acquisition posture of the preset camera at each acquisition angle can be accurately and effectively determined. Secondly, the acquisition angles of different categories of automobile parts and the corresponding acquisition postures of the preset cameras are recorded to accurately and effectively construct the monitoring layout guidance strategy library, and category identifications are generated based on the category information of automobile parts to effectively associate the monitoring layout guidance with the category identification. Finally, the access configuration of the monitoring layout guidance is completed according to the association results, which provides accurate and reliable configuration parameters for image acquisition of different categories of automobile parts, thereby improving the reliability of multi-angle component image acquisition of automobile parts.

[0063] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be achieved and obtained through the structures specifically pointed out in this application document.

[0064] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0066] Figure 1 This is a flow chart of a method for identifying and locating automobile parts based on machine vision in an embodiment of the present invention;

[0067] Figure 2 This is a flowchart of step 1 in a method for identifying and locating automotive parts based on machine vision in an embodiment of the present invention;

[0068] Figure 3 This is a structural diagram of an automobile parts recognition and positioning system based on machine vision in an embodiment of the present invention. DETAILED DESCRIPTION

[0069] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0070] Example 1:

[0071] This embodiment provides a method for identifying and locating automobile parts based on machine vision, such as Figure 1 Shown, including:

[0072] Step 1: Capture multi-angle images of auto parts using a preset camera to obtain multi-angle component images, and perform standardization on the multi-angle component images;

[0073] Step 2: Preprocess the standardized multi-angle component images, and extract component representation features in the multi-angle component images based on the preprocessing results;

[0074] Step 3: Identify and locate the automobile parts based on the component characterization features, and obtain the specific position coordinates of the automobile parts based on the identification and positioning results.

[0075] In this embodiment, the preset camera is a tool set in advance and used to capture images of automobile parts.

[0076] In this embodiment, multi-angle image acquisition refers to capturing images of the same automotive component from multiple different angles using a preset camera, wherein the multi-angle component image is the result obtained after image acquisition from multiple angles.

[0077] In this embodiment, the standardization process refers to converting different coordinate systems corresponding to the multi-angle images into images under the same coordinate system, thereby facilitating preprocessing and feature extraction of the multi-angle images.

[0078] In this embodiment, pre-processing refers to performing grayscale processing and noise removal on the multi-angle component images.

[0079] In this embodiment, the component representation features refer to the appearance features of the automobile component reflected by the multi-angle component images, including the shape, texture, and color of the automobile component.

[0080] In this embodiment, identification and positioning refers to determining the category of the automobile component and its location in the automobile, wherein the specific location coordinates are the specific location information of the automobile component.

[0081] The working principle and beneficial effects of the above technical solution are: multi-angle images of automobile parts are collected by pre-setting cameras, and the collected multi-angle component images are standardized, so as to facilitate effective preprocessing of the multi-angle component images; secondly, component representation features are extracted from the pre-processed multi-angle component images, so as to facilitate identification and positioning of automobile parts based on the extracted component representation features; finally, by analyzing the component representation features, effective identification and positioning of automobile parts are achieved, thereby improving the accuracy and effectiveness of determining the specific position coordinates of automobile parts, and ensuring the accuracy and efficiency of identification and positioning of automobile parts.

[0082] Example 2:

[0083] Based on Example 1, this embodiment provides a method for identifying and locating automobile parts based on machine vision. In step 1, before multi-angle image acquisition of automobile parts based on a preset camera is performed, the method includes:

[0084] Obtain image acquisition requirements for automotive parts, and based on the image acquisition requirements, determine acquisition angles for different types of automotive parts and preset camera acquisition postures at each acquisition angle;

[0085] Record the acquisition angles of different types of automotive parts and the corresponding acquisition postures of preset cameras, and generate a monitoring layout guidance strategy library for preset cameras based on the recording results;

[0086] Generate category identifiers based on category information of automobile parts, and determine target addresses of monitoring layout guidelines for different categories of automobile parts in a monitoring layout guide strategy library;

[0087] Based on the target address, a jump link is configured for different category identifiers, and based on the jump link, the access configuration of the category identifier and the corresponding monitoring layout guide in the monitoring layout guide policy library is completed.

[0088] In this embodiment, the image acquisition requirements are known in advance and are used to characterize the acquisition angle and acquisition posture of the automobile parts.

[0089] In this embodiment, the monitoring layout guidance strategy library refers to configuration schemes of preset cameras of all categories obtained by recording the acquisition angles and acquisition postures of different categories of automobile parts.

[0090] In this embodiment, the category identifier is a marking symbol used to mark different categories of automotive parts, and each category corresponds to a category identifier.

[0091] In this embodiment, the target address refers to the specific storage address corresponding to different monitoring layout guidelines when they are stored in the monitoring layout guideline policy library.

[0092] In this embodiment, the jump link refers to a hyperlink, that is, the user can directly jump to the corresponding monitoring layout guide through the category identifier.

[0093] The working principle and beneficial effects of the above technical solution are: by analyzing the image acquisition requirements of automobile parts, the acquisition angles of different categories of automobile parts and the acquisition posture of the preset camera at each acquisition angle are accurately and effectively determined; secondly, the acquisition angles of different categories of automobile parts and the corresponding acquisition postures of the preset cameras are recorded to accurately and effectively construct a monitoring layout guidance strategy library, and generate category identifications based on the category information of automobile parts to effectively associate the monitoring layout guidance with the category identification, and finally complete the access configuration of the monitoring layout guidance based on the association results, providing accurate and reliable configuration parameters for image acquisition of different categories of automobile parts, thereby improving the reliability of multi-angle component image acquisition of automobile parts.

[0094] Example 3:

[0095] Based on Example 1, this embodiment provides a method for identifying and locating automobile parts based on machine vision, such as Figure 2 As shown, in step 1, multi-angle image acquisition of automobile parts is performed based on a preset camera, including:

[0096] Step 101: Acquire acquisition angles for automobile parts, and determine the execution order of different acquisition angles based on acquisition requirements;

[0097] Step 102: At the same time, based on the acquisition requirements, the adjustable range of the preset camera at each acquisition angle is determined, and the target shooting frequency of the preset camera at each acquisition posture within the adjustable range is determined;

[0098] Step 103: Based on the target shooting frequency, the preset camera is controlled to capture images of the automobile parts at corresponding acquisition angles according to the execution order to obtain multi-angle component images.

[0099] In this embodiment, the collection requirements are set in advance and are used to represent the collection standards for automobile parts, including the order corresponding to the collection angles.

[0100] In this embodiment, the adjustable range refers to the range in which the preset camera can adjust the angle at different acquisition angles.

[0101] In this embodiment, the target shooting frequency refers to the number of images that the preset camera needs to capture in each acquisition posture.

[0102] The working principle and beneficial effects of the above technical solution are: by determining the acquisition angle of automobile parts, the target shooting frequency at each acquisition angle is determined according to the acquisition requirements, which provides effective limiting conditions for effective image acquisition. Finally, accurate and effective acquisition of automobile parts is achieved based on the target shooting frequency and acquisition angle, ensuring the reliability and comprehensiveness of the final multi-angle component images.

[0103] Example 4:

[0104] Based on Example 3, this embodiment provides a method for identifying and locating automobile parts based on machine vision, which obtains multi-angle part images, including:

[0105] Acquire the obtained multi-angle component images and extract the shooting sources of the multi-angle component images;

[0106] Generate location tags for component images at different angles based on the shooting source, and perform location marking on the corresponding component images at different angles based on the location tags;

[0107] The multi-angle component images after position marking are classified by angle, and based on the classification results, the transmission queues of the automobile component images acquired at different angles are allocated;

[0108] Based on the allocation results, the images of automobile parts under different acquisition angles are transmitted to the management terminal in turn.

[0109] In this embodiment, the shooting source refers to the acquisition position corresponding to the multi-angle component image, etc.

[0110] In this embodiment, the position label is a marking symbol used to distinguish component images at different angles, and component images at different angles correspond to different position labels.

[0111] In this embodiment, the transmission queue is a carrier for transmitting the images of automobile parts at different acquisition angles, and is used to transmit the images of automobile parts to the management terminal.

[0112] The working principle and beneficial effects of the above technical solution are: by determining the shooting source corresponding to the multi-angle component image, effective angle classification of the multi-angle component image is achieved according to the shooting source; finally, according to the angle classification results, a transmission queue is allocated to the automobile component images at different acquisition angles, and the automobile component images at different acquisition angles are effectively transmitted to the management terminal through the transmission queue, thereby providing reliable data support for automobile component identification and positioning.

[0113] Example 5:

[0114] Based on Example 1, this embodiment provides a method for identifying and locating automobile parts based on machine vision. In step 1, multi-angle component images are standardized, including:

[0115] Obtain the imaging coordinate system of the multi-angle component image at the corresponding acquisition angle, and at the same time, determine the target unified coordinate system for the multi-angle component image;

[0116] Determine the relative relationship between the imaging coordinate system at different acquisition angles and the target unified coordinate system, and determine the difference change scheme of the imaging coordinate system at different acquisition angles based on the relative relationship;

[0117] Based on the difference change scheme, the imaging coordinate systems under different acquisition angles are transformed, and a unified component image under the target unified coordinate system is obtained based on the coordinate transformation results;

[0118] The unified component image is sampled, and the sampling result is quality evaluated based on the image quality evaluation index. When the quality evaluation result does not meet the preset requirements, the unified component image is adjusted in multiple dimensions until the preset requirements are met.

[0119] In this embodiment, the imaging coordinate system refers to the imaging rules of the multi-angle component image at the corresponding acquisition angles.

[0120] In this embodiment, the target unified coordinate system refers to a coordinate system that unifies the scales and specifications of imaging coordinate systems at different acquisition angles, and is used to unify multi-angle component images, thereby facilitating effective recognition of multi-angle component images.

[0121] In this embodiment, the difference change solution refers to a specific method of converting imaging coordinate systems under different acquisition angles into a target unified coordinate system.

[0122] In this embodiment, the unified component image refers to a final image obtained by performing coordinate system unification on component images acquired at different acquisition angles.

[0123] In this embodiment, the image quality evaluation indexes are set in advance, including image clarity, image resolution, image color, etc.

[0124] The working principle and beneficial effects of the above technical solution are: by determining the relative relationship between the imaging coordinate system of the multi-angle component image at the corresponding acquisition angle and the target unified coordinate system, the imaging coordinate system at different acquisition angles is transformed according to the relative relationship, thereby realizing standardized processing of the multi-angle component image, which provides convenience for the preprocessing of the multi-angle component image.

[0125] Example 6:

[0126] Based on Example 1, this embodiment provides a method for identifying and locating automobile parts based on machine vision. In step 2, the standardized multi-angle component images are preprocessed, including:

[0127] Read the standardized multi-angle component images, and perform multi-dimensional image quality assessment on each angle component image according to the image quality assessment dimension;

[0128] Output the quality assessment result of each angular component image according to the assessment result, read the quality assessment result, and determine the assessment value corresponding to each assessment dimension in each angular component image;

[0129] Sort the evaluation values ​​corresponding to each evaluation dimension in each angle component image in ascending order, and generate a dimension processing priority for each angle component image according to the sorting result;

[0130] Obtaining the dimension weight of the quality assessment dimension, and adapting the dimension weight of the quality assessment dimension to the dimension processing priority corresponding to each angular component image, respectively, to obtain the degree of adaptation between the dimension weight of the quality assessment dimension and the dimension processing priority corresponding to each angular component image;

[0131] Sorting the degree of fit between the dimension weights of the quality assessment dimensions and the dimension processing priorities corresponding to each angular component image from large to small, and obtaining the image processing priorities of the plurality of angular component images according to the sorting results;

[0132] Generate a first preprocessing strategy based on the image processing priorities of the plurality of angular component images, and generate a second preprocessing strategy based on the dimensional processing priority corresponding to each angular component image;

[0133] The second preprocessing strategy is integrated into the first preprocessing strategy to generate a comprehensive preprocessing strategy for preprocessing the standardized multi-angle component image;

[0134] The standardized multi-angle component images are preprocessed based on a comprehensive preprocessing strategy.

[0135] In this embodiment, the image quality assessment dimension refers to items for quality assessment of multi-angle component images, including resolution, clarity, and color.

[0136] In this embodiment, the evaluation value is used to represent the quality evaluation result corresponding to each evaluation dimension in each angle component image. A larger value indicates a better quality of the angle component image.

[0137] In this embodiment, the dimension processing priority is determined according to the order of evaluation values ​​from small to large, and is used to limit the order of processing component images at different angles.

[0138] In this embodiment, the dimension weight is used to represent the importance of different quality assessment dimensions when performing quality assessment.

[0139] In this embodiment, the degree of adaptation is the degree of correspondence between the dimension weight used to characterize the quality assessment dimension and the dimension processing priority corresponding to each angular component image.

[0140] In this embodiment, the first pre-processing strategy is generated according to the image processing priorities of the component images at multiple angles.

[0141] In this embodiment, the second pre-processing strategy is generated according to the dimension processing priority.

[0142] In this embodiment, fusing the second preprocessing strategy into the first preprocessing strategy refers to associating and coupling the preprocessing strategies corresponding to different dimensional processing of the same image, thereby achieving effective preprocessing of the image.

[0143] In this embodiment, the comprehensive preprocessing strategy refers to a final preprocessing solution corresponding to component images at different angles obtained after fusing the second preprocessing strategy into the first preprocessing strategy.

[0144] The working principle and beneficial effects of the above technical solution are: by analyzing the multi-angle component images after standardization, the image quality assessment dimension is determined, so that the multi-dimensional image quality assessment of the multi-angle component images is carried out through the image quality assessment dimension, and the dimension processing priority and image processing priority of each angle component image are determined according to the multi-dimensional image quality assessment results, so as to facilitate the orderly preprocessing of the multi-angle component images. Finally, a first preprocessing strategy is generated according to the image processing priority of the multiple angle component images. At the same time, a second preprocessing strategy is generated according to the dimension processing priority corresponding to each angle component image, and the second preprocessing strategy is integrated into the first preprocessing strategy to achieve accurate and effective formulation of the final preprocessing strategy for the multi-angle component images, and to achieve accurate and reliable preprocessing of the standardized multi-angle component images according to the formulated preprocessing strategy, thereby ensuring the reliability of the multi-angle component images and providing convenience for improving the identification and positioning of automotive parts.

[0145] Example 7:

[0146] Based on Example 1, this embodiment provides a method for identifying and locating automobile parts based on machine vision. In step 2, component representation features are extracted from multi-angle component images based on preprocessing results, including:

[0147] Acquire the pre-processed multi-angle component image, analyze the multi-angle component image, and determine the depth information of the multi-angle component image;

[0148] Determine a target distance of the automobile part relative to a preset camera based on the depth information, determine an imaging size of the automobile part based on the target distance, and lock the imaging size;

[0149] Based on the locking results, pixel points of the multi-angle component images under corresponding imaging sizes are traversed respectively, and local features in the multi-angle component images are determined respectively based on the traversal results;

[0150] Obtain corresponding feature points based on local features, and perform regional segmentation on multi-angle component images based on the feature points;

[0151] Extract sensitive image areas of auto parts based on the region segmentation results, perform edge detection on the sensitive image areas, and determine the contour features of auto parts;

[0152] At the same time, pixel analysis is performed on sensitive image areas to determine the phenomenon representation features of automobile parts, and the contour features of automobile parts are summarized with the phenomenon representation features of automobile parts to obtain the component representation features in multi-angle component images.

[0153] In this embodiment, the depth information refers to the distance information of each automobile component relative to a preset camera in the multi-angle component image.

[0154] In this embodiment, the target distance is used to represent the distance between the automobile parts and the preset camera.

[0155] In this embodiment, the imaging size refers to the scale information of the automobile part image, including the length and width of the automobile part image.

[0156] In this embodiment, the local features refer to parameter information in the multi-angle component image obtained by performing pixel traversal on the multi-angle component image, which can characterize the shape and structure of the automobile parts.

[0157] In this embodiment, the feature points refer to position points that can characterize the specific structure and posture of automobile parts.

[0158] In this embodiment, the sensitive image area refers to the main image area containing automobile parts in the multi-angle component image.

[0159] In this embodiment, the contour feature refers to the specific contour shape of the automobile part.

[0160] In this embodiment, the phenomenon representation features refer to the color, texture, and shape of automobile parts.

[0161] The working principle and beneficial effects of the above technical solution are: by parsing the pre-processed multi-angle component image, the depth information of the multi-angle component image can be accurately and effectively determined, so that the target distance of the automobile component relative to the preset camera can be locked according to the depth information, and the imaging size of the automobile component can be determined according to the target distance. Secondly, the pixel points of the multi-angle component image are traversed according to the imaging size to determine the local features in the multi-angle component image, so that the multi-angle component image can be regionally segmented according to the local features, which provides convenience for automobile component identification. Finally, the sensitive image area of ​​the automobile component is extracted according to the regional segmentation result, and the edge detection is performed on the sensitive image area to lock the contour features of the automobile component. Finally, the obtained contour features are summarized with the phenomenon representation features of the automobile component to achieve accurate and effective determination of the component representation features in the multi-angle component image, thereby ensuring the accuracy and efficiency of automobile component identification and positioning.

[0162] Example 8:

[0163] Based on Example 1, this embodiment provides a method for identifying and locating automobile parts based on machine vision. In step 3, the automobile parts are identified and located based on the component characterization features, and the specific position coordinates of the automobile parts are obtained based on the identification and positioning results, including:

[0164] Obtaining the obtained component characterization features, matching the component characterization features with preset component features, and determining the part category of the automotive component based on the matching results;

[0165] Extract the functional attributes of part categories and determine the distribution range of automotive parts based on the functional attributes;

[0166] Retrieving a local distribution map of automotive components based on the distribution range, and comparing the component characterization features with the known components in the local distribution map of automotive components for similarity;

[0167] Based on the similarity comparison result, the target position of the automobile parts in the local distribution map of automobile components is determined, and the coordinate transformation of the target position is performed based on the position relationship between the local distribution map of automobile components and the three-dimensional automobile coordinate position to obtain the specific position coordinates of the automobile parts.

[0168] In this embodiment, the preset component features are known in advance and are used to characterize the specific structure and morphological characteristics of each component.

[0169] In this embodiment, the functional attribute refers to the specific role played by the part category in the automobile structure, such as a screw for fastening.

[0170] In this embodiment, the automobile component local distribution map is a diagram used to represent the distribution of various automobile components in the automobile.

[0171] In this embodiment, the target position is used to represent the corresponding position of the automobile component in the automobile device local distribution map.

[0172] The working principle and beneficial effects of the above technical solution are: by matching the component characterization features with the preset component features, the part category of the automobile part is accurately and effectively determined; secondly, the functional attributes of the part category are extracted to determine the distribution range of the automobile part according to the functional attributes; finally, the local distribution map of the automobile device is retrieved according to the distribution range, and the specific position coordinates of the automobile part are accurately and effectively determined by comparing the similarity between the component characterization features and the known devices in the local distribution map of the automobile device, thereby improving the accuracy and reliability of the identification and positioning of automobile parts.

[0173] Example 9:

[0174] This embodiment provides an automobile parts recognition and positioning system based on machine vision, such as Figure 3 Shown, including:

[0175] Image acquisition and standardization module, used to acquire multi-angle images of automotive parts based on preset cameras, obtain multi-angle component images, and perform standardization on the multi-angle component images;

[0176] A feature extraction module is used to preprocess the standardized multi-angle component images and extract component representation features in the multi-angle component images based on the preprocessing results;

[0177] The identification and positioning module is used to identify and locate automobile parts based on the component characterization features, and obtain the specific position coordinates of the automobile parts based on the identification and positioning results.

[0178] The working principle and beneficial effects of the above technical solution are: multi-angle images of automobile parts are collected by pre-setting cameras, and the collected multi-angle component images are standardized, so as to facilitate effective preprocessing of the multi-angle component images; secondly, component representation features are extracted from the pre-processed multi-angle component images, so as to facilitate identification and positioning of automobile parts based on the extracted component representation features; finally, by analyzing the component representation features, effective identification and positioning of automobile parts are achieved, thereby improving the accuracy and effectiveness of determining the specific position coordinates of automobile parts, and ensuring the accuracy and efficiency of identification and positioning of automobile parts.

[0179] Example 10:

[0180] Based on Example 9, this embodiment provides a method for identifying and locating automobile parts based on machine vision, and an image acquisition and standardization module, including:

[0181] A configuration unit is used to obtain image acquisition requirements for automobile parts, and based on the image acquisition requirements, determine acquisition angles for different types of automobile parts and preset camera acquisition postures at each acquisition angle;

[0182] A monitoring layout guidance construction unit is used to record the acquisition angles of different types of automotive parts and the corresponding acquisition postures of preset cameras, and generate a monitoring layout guidance strategy library for preset cameras based on the recording results;

[0183] Monitoring layout guidance reference units, including:

[0184] Generate category identifiers based on category information of automobile parts, and determine target addresses of monitoring layout guidelines for different categories of automobile parts in a monitoring layout guide strategy library;

[0185] Based on the target address, a jump link is configured for different category identifiers, and based on the jump link, the access configuration of the category identifier and the corresponding monitoring layout guide in the monitoring layout guide policy library is completed.

[0186] The working principle and beneficial effects of the above technical solution are: by analyzing the image acquisition requirements of automobile parts, the acquisition angles of different categories of automobile parts and the acquisition posture of the preset camera at each acquisition angle are accurately and effectively determined; secondly, the acquisition angles of different categories of automobile parts and the corresponding acquisition postures of the preset cameras are recorded to accurately and effectively construct a monitoring layout guidance strategy library, and generate category identifications based on the category information of automobile parts to effectively associate the monitoring layout guidance with the category identification, and finally complete the access configuration of the monitoring layout guidance based on the association results, providing accurate and reliable configuration parameters for image acquisition of different categories of automobile parts, thereby improving the reliability of multi-angle component image acquisition of automobile parts.

[0187] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A method for identifying and locating automobile parts based on machine vision, characterized in that: include: Step 1: Based on the preset camera, multi-angle image acquisition is performed on the automobile parts to obtain multi-angle component images, and the multi-angle component images are standardized; Step 2: preprocessing the standardized multi-angle component images, and extracting component representation features in the multi-angle component images based on the preprocessing results; Step 3: Identify and locate automobile parts based on the component characterization features, and obtain the specific position coordinates of the automobile parts based on the identification and positioning results.

2. The method for identifying and locating automobile parts based on machine vision according to claim 1, characterized in that: In step 1, before multi-angle image acquisition of automobile parts based on a preset camera, the following steps are performed: Obtain image acquisition requirements for automobile parts, and determine acquisition angles for different types of automobile parts and acquisition postures of preset cameras at each acquisition angle based on the image acquisition requirements; Record the acquisition angles of different types of automotive parts and the corresponding acquisition postures of preset cameras, and generate a monitoring layout guidance strategy library for preset cameras based on the recording results; Generate a category identifier based on the category information of the automobile parts, and determine the target address of the monitoring layout guidance of different categories of automobile parts in the monitoring layout guidance strategy library; Based on the target address, a jump link is configured for different category identifiers, and based on the jump link, the access configuration of the category identifier and the corresponding monitoring layout guide in the monitoring layout guide policy library is completed.

3. The method for identifying and locating automobile parts based on machine vision according to claim 1, characterized in that: In step 1, multi-angle image acquisition of automobile parts is performed based on a preset camera, including: Obtain the acquisition angles of automobile parts and determine the execution order between different acquisition angles based on acquisition requirements; At the same time, based on the acquisition requirements, the adjustable range of the preset camera at each acquisition angle is determined, and the target shooting frequency of the preset camera at each acquisition posture within the adjustable range is determined; Based on the target shooting frequency, the preset camera is controlled to collect images of automobile parts at corresponding collection angles according to the execution order to obtain multi-angle component images.

4. The method for identifying and locating automobile parts based on machine vision according to claim 3, characterized in that: Get multi-angle component images, including: Acquire the obtained multi-angle component images, and extract the shooting sources of the multi-angle component images; Generate position tags of component images at different angles based on the shooting source, and mark the positions of corresponding component images at different angles based on the position tags; The multi-angle component images after position marking are classified by angle, and the transmission queues are allocated to the automobile component images at different acquisition angles based on the classification results; Based on the allocation results, the images of automobile parts under different acquisition angles are transmitted to the management terminal in turn.

5. The method for identifying and locating automobile parts based on machine vision according to claim 1, characterized in that: In step 1, the multi-angle component images are standardized, including: Acquire the imaging coordinate system of the multi-angle component image at the corresponding acquisition angle, and at the same time, determine the target unified coordinate system for the multi-angle component image; Determine the relative relationship between the imaging coordinate system at different acquisition angles and the target unified coordinate system, and determine the difference change scheme of the imaging coordinate system at different acquisition angles based on the relative relationship; Based on the difference change scheme, coordinate transformation is performed on the imaging coordinate systems under different acquisition angles, and a unified component image under the target unified coordinate system is obtained based on the coordinate transformation result; The unified component image is sampled, and the sampling result is quality evaluated based on the image quality evaluation index. When the quality evaluation result does not meet the preset requirements, the unified component image is adjusted in multiple dimensions until the preset requirements are met.

6. The method for identifying and locating automobile parts based on machine vision according to claim 1, characterized in that: In step 2, the standardized multi-angle component images are preprocessed, including: Read the standardized multi-angle component images, and perform multi-dimensional image quality assessment on each angle component image according to the image quality assessment dimension; Output the quality assessment result of each angle component image according to the assessment result, read the quality assessment result, and determine the assessment value corresponding to each assessment dimension in each angle component image; The evaluation values ​​corresponding to each evaluation dimension in each angle component image are sorted in order from small to large, and a dimension processing priority is generated for each angle component image according to the sorting result; Obtaining the dimension weight of the quality assessment dimension, and adapting the dimension weight of the quality assessment dimension to the dimension processing priority corresponding to each angular component image, respectively, to obtain the degree of adaptation of the dimension weight of the quality assessment dimension to the dimension processing priority corresponding to each angular component image; Sorting the degree of fit between the dimension weight of the quality assessment dimension and the dimension processing priority corresponding to each angle component image from large to small, and obtaining the image processing priority of multiple angle component images according to the sorting result; Generate a first preprocessing strategy according to the image processing priorities of the plurality of angle component images, and generate a second preprocessing strategy according to the dimensional processing priority corresponding to each angle component image; The second preprocessing strategy is integrated into the first preprocessing strategy to generate a comprehensive preprocessing strategy for preprocessing the standardized multi-angle component images; The standardized multi-angle component images are preprocessed based on a comprehensive preprocessing strategy.

7. The method for identifying and locating automobile parts based on machine vision according to claim 1, characterized in that: In step 2, component representation features in multi-angle component images are extracted based on the preprocessing results, including: Acquire the preprocessed multi-angle component image, analyze the multi-angle component image, and determine the depth information of the multi-angle component image; Determine a target distance of the automobile part relative to a preset camera based on the depth information, determine an imaging size of the automobile part based on the target distance, and lock the imaging size; Based on the locking results, pixel points of the multi-angle component images under the corresponding imaging size are traversed respectively, and local features in the multi-angle component images are determined respectively based on the traversal results; Obtain corresponding feature points based on local features, and perform regional segmentation on multi-angle component images based on the feature points; Extract sensitive image areas of automobile parts based on the region segmentation results, perform edge detection on the sensitive image areas, and determine the contour features of automobile parts; At the same time, pixel analysis is performed on sensitive image areas to determine the phenomenon characterization features of automobile parts, and the contour features of automobile parts are summarized with the phenomenon characterization features of automobile parts to obtain the component characterization features in multi-angle component images.

8. The method for identifying and locating automobile parts based on machine vision according to claim 1, characterized in that: In step 3, the automobile parts are identified and located based on the component characterization features, and the specific position coordinates of the automobile parts are obtained based on the identification and positioning results, including: Acquire the obtained component characterization features, match the component characterization features with preset component features, and determine the part category of the automobile component based on the matching result; Extract the functional attributes of the part categories and determine the distribution range of automobile parts based on the functional attributes; Retrieving a local distribution map of automobile components based on the distribution range, and comparing the component characterization features with each known component in the local distribution map of automobile components for similarity; The target position of the automobile parts in the local distribution map of automobile components is determined based on the similarity comparison result, and the target position is transformed based on the position relationship between the local distribution map of automobile components and the three-dimensional automobile coordinates to obtain the specific position coordinates of the automobile parts.

9. An automobile parts recognition and positioning system based on machine vision, characterized in that: include: An image acquisition and standardization module is used to acquire multi-angle images of automobile parts based on a preset camera, obtain multi-angle component images, and perform standardization on the multi-angle component images; A feature extraction module is used to preprocess the standardized multi-angle component images and extract component representation features in the multi-angle component images based on the preprocessing results; The identification and positioning module is used to identify and locate automobile parts based on the component characterization features, and obtain the specific position coordinates of the automobile parts based on the identification and positioning results.

10. The machine vision-based automobile parts identification and positioning system according to claim 9, characterized in that: Image acquisition and standardization module, including: A configuration unit, used to obtain image acquisition requirements for automobile parts, and determine acquisition angles for different types of automobile parts and acquisition postures of preset cameras at each acquisition angle based on the image acquisition requirements; A monitoring layout guidance construction unit, used to record the acquisition angles of different categories of automobile parts and the corresponding acquisition postures of preset cameras, and generate a monitoring layout guidance strategy library of preset cameras based on the recording results; Monitor layout guide reference units, including: Generate a category identifier based on the category information of the automobile parts, and determine the target address of the monitoring layout guidance of different categories of automobile parts in the monitoring layout guidance strategy library; Based on the target address, a jump link is configured for different category identifiers, and based on the jump link, the access configuration of the category identifier and the corresponding monitoring layout guide in the monitoring layout guide policy library is completed.

Images