Multi-type workpiece classification system based on knowledge and data

Abstract

A knowledge and data-based multi-type workpiece classification system disclosed by the present invention comprises an image acquisition system and an industrial computer system, the image acquisition system is in signal connection with the industrial computer system, and the industrial computer system is in signal connection with an upper computer. The image acquisition system can be mounted or dismounted at a detection station of a production line, the image acquisition system is provided with a sensor, and the acquisition function of the image acquisition system is triggered through the sensor. The invention relates to the technical field of intelligent machine vision, in particular to a multi-type workpiece classification system based on knowledge and data, which can perform workpiece image data acquisition through an industrial camera, realizes workpiece type identification by using workpiece feature information and a deep learning method, improves the identification accuracy of sorting, and improves the classification efficiency. And the actual production requirements are met.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent machine vision, and in particular is a multi-type workpiece classification system based on knowledge and data. Background Art

[0002] With the vigorous development of the automobile industry, sorting auto parts in the production line has become an issue that cannot be ignored. Auto parts are of various types, complex shapes, sizes, and materials. Factories often need to sort these parts for assembly or logistics packaging.

[0003] Currently, the workpiece sorting relies heavily on manual labor, and the sorting quality is difficult to guarantee. The concept of labor load rate is introduced to describe the working conditions of workers. The labor load rate is equal to the average value of the daily load curve divided by the maximum value of the daily load curve multiplied by 100%. The higher the labor load rate of workers, the more difficult it is for them to do their current work, and the workers will become tired and their work efficiency will decrease. The high cost and efficiency of manual sorting will also affect the health of workers. Long-term labor can easily cause visual fatigue, which will cause the reliability of sorting results to decrease to a certain extent after a long working time.

[0004] With the rapid development of machine vision and artificial intelligence technology, more and more manufacturing companies are beginning to choose to use machine vision to sort automotive parts to replace efficient and low-cost manual sorting. Compared with manual sorting, intelligent sorting based on machine vision has the advantages of high efficiency, good consistency and low cost.

[0005] However, in actual application, due to factors such as different degrees of difference in sorting targets, unbalanced target samples, and multiple target categories, the recognition accuracy is low and it is difficult to meet actual production needs. Therefore, improvements need to be made to the existing problems. Summary of the invention

[0006] In view of the above situation, in order to make up for the above-mentioned existing defects, the present invention provides a multi-type workpiece classification system based on knowledge and data. The system can collect workpiece image data through industrial cameras, use workpiece feature information and deep learning methods to realize workpiece type recognition, thereby improving the recognition accuracy of sorting and meeting actual production needs.

[0007] The present invention proposes a knowledge and data-based multi-type workpiece classification system, including an image acquisition system and an industrial computer system. The image acquisition system is signal-connected to the industrial computer system, and the industrial computer system is signal-connected to a host computer. The image acquisition system can be installed or removed at a detection station on a production line. The image acquisition system is provided with a sensor, and the acquisition function of the image acquisition system is triggered by the sensor. The image acquisition system is signal-connected to the industrial computer system.

[0008] The image acquisition system further includes a light source, a light source controller, filters of different wavelengths, a lens, a laser generator and an industrial camera.

[0009] Among them, the industrial computer system includes a communication module, an image processing module, a data annotation module and a data storage module. The communication module contains a camera control module and a light source control module. The communication module is connected to the image acquisition system signal and is used to adjust the light source brightness and control the acquisition of the industrial camera through the communication module.

[0010] Furthermore, the image processing module uses traditional image processing methods and deep learning methods. The image processing module can extract and classify features of the workpiece and transmit the processed data to the host computer. The image processing module can also store the processed data locally or in the cloud.

[0011] Furthermore, the image processing module comprises an offline template library and an online detection submodule, through which the workpiece can be detected online, and compared, matched and scored with the workpiece stored in the offline template library.

[0012] Furthermore, the image processing module is connected to the host computer signal, and the data annotation module has an interactive function. When the image processing module detects and processes a new workpiece, a new category annotation can be manually performed through the data annotation module for iterative training of the model.

[0013] The beneficial effects achieved by the present invention using the above structure are as follows: The advantages of the multi-type artifact classification system based on knowledge and data of the present invention are:

[0014] 1. This system can perform online detection of artifacts through offline template library and online detection submodule, and compare, match and score artifacts stored in the offline template library.

[0015] 2. This system can manually label parts with new categories through the data labeling module for iterative training of the model.

[0016] 3. This system cooperates with the parts detection algorithm to realize the model recognition of automobile parts, send alarm signals to the host computer in real time, prompt manual intervention, avoid additional costs caused by incorrect assembly of parts, and greatly improve the level of factory intelligence.

[0017] 4. This system does not require manual inventory and can identify and classify auto parts in real time with high precision, saving manpower, material resources and time, reducing costs and improving subsequent processing and assembly efficiency and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] 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:

[0019] Figure 1 It is a schematic diagram of the principle structure of the present invention based on knowledge and data. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0021] It should be noted that the words "front", "rear", "left", "right", "up" and "down" used in the following description refer to directions in the drawings, and the words "inside" and "outside" refer to directions toward or away from the geometric center of a specific component, respectively.

[0022] like Figure 1As shown, the technical solution adopted by the present invention is as follows: a multi-type workpiece classification system based on knowledge and data proposed by the present invention includes an image acquisition system and an industrial computer system. The image acquisition system is connected to the industrial computer system by signal, and the industrial computer system is connected to the host computer by signal. The image acquisition system can be installed or disassembled at the inspection station of the production line. The image acquisition system has a sensor, and the acquisition function of the image acquisition system is triggered by the sensor. The image acquisition system is connected to the industrial computer system by signal. The image acquisition system also includes a light source, a light source controller, filters of different wavelengths, a lens, a laser generator and an industrial camera. The industrial computer system includes a communication module, an image processing module, a data annotation module and a data storage module. The communication module includes a camera control module and a light source control module. The communication module is connected to the image acquisition system by signal, and is used to adjust the brightness of the light source and control the acquisition of the industrial camera. The image processing module uses traditional image processing methods and deep learning methods. The image processing module can extract and classify the features of the workpiece, and can transmit the processed data to the host computer.

[0023] The image processing module can also store the processed data locally or in the cloud. The image processing module includes an offline template library and an online detection submodule. The offline template library and the online detection submodule can perform online detection of artifacts, and compare, match and score them with the artifacts stored in the offline template library.

[0024] The image processing module is connected to the host computer signal, and the data labeling module has an interactive function. When the image processing module detects and processes new workpieces, new category labels can be manually performed through the data labeling module for iterative training of the model.

[0025] When used specifically, the image acquisition system is installed at the production line inspection station of automobile tires to sample tire surface data. When classifying tires, the industrial computer system controls the line structured light sensor of the camera acquisition system through the camera control module to acquire point cloud images. The image acquisition system returns the acquired point cloud images to the industrial computer and processes them through the image processing module. By setting up the offline template library and online detection submodule in the image processing module, the offline template library can be compared with the offline template library for matching and scoring according to the different tread patterns of the tires. After obtaining the highest matching score, the result is returned to the host computer, and the point cloud image and result data are stored locally or in the cloud.

[0026] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, material or apparatus including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, material or apparatus.

[0027] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-type artifact classification system based on knowledge and data, characterized by: It includes an image acquisition system and an industrial computer system. The image acquisition system is connected to the industrial computer system by signal. The industrial computer system is connected to a host computer by signal. The image acquisition system can be installed or removed at a detection station of a production line. The image acquisition system is provided with a sensor. The acquisition function of the image acquisition system is triggered by the sensor. The image acquisition system is connected to the industrial computer system by signal.

2. The multi-type artifact classification system based on knowledge and data according to claim 1 is characterized by: The image acquisition system also includes a light source, a light source controller, filters of different wavelengths, a lens, a laser generator and an industrial camera.

3. The multi-type artifact classification system based on knowledge and data according to claim 2 is characterized by: The industrial computer system includes a communication module, an image processing module, a data annotation module and a data storage module. The communication module includes a camera control module and a light source control module. The communication module is connected to the image acquisition system signal and is used to adjust the light source brightness and control the acquisition of the industrial camera through the communication module.

4. The knowledge and data-based multi-type artifact classification system according to claim 3, characterized in that: The image processing module uses traditional image processing methods and deep learning methods. The image processing module can extract and classify features of workpieces and transmit the processed data to a host computer. The image processing module can also store the processed data locally or in the cloud.

5. The knowledge and data-based multi-type artifact classification system according to claim 4, characterized in that: The image processing module comprises an offline template library and an online detection submodule, through which the workpiece can be detected online, and compared, matched and scored with the workpiece stored in the offline template library.

6. The knowledge and data-based multi-type artifact classification system according to claim 5, characterized in that: The image processing module is connected to the host computer signal, and the data annotation module has an interactive function. When the image processing module detects and processes a new workpiece, a new category annotation can be manually performed through the data annotation module for iterative training of the model.

Images