Intelligent surface detection method and system for medium plate

Abstract

The invention discloses an intelligent surface detection method for a medium plate, and relates to the technical field of steel plate inspection, and the method comprises the following steps: collecting image data of a to-be-detected steel plate; dividing the acquired steel plate image data into a first detection area and a second detection area; and presetting first detection precision and second detection precision of the detection model, performing defect detection on the image data of the first detection area by adopting the first detection precision, and performing defect detection on the second detection area by adopting the second detection precision. According to the method, the head steel plate and the tail steel plate in medium and thick plate production can be rapidly recognized, meanwhile, the recognition precision of key areas with high defects is enhanced, the false detection rate of abnormal steel billets can be effectively reduced, and on the basis of improving the defect recognition accuracy and guaranteeing the surface quality of the steel plates, the detection process is further optimized, and the overall working efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of steel plate inspection technology, and in particular to an intelligent surface inspection method and system for medium and heavy plates. Background Technology

[0002] In the production of medium and heavy plates, due to the objective factor of a fast production pace, the existing intelligent surface inspection algorithm equipment for medium and heavy plates has obvious technical limitations. The core problem is that it cannot accurately identify the steel plate head and tail plates and abnormal billets in the production process, resulting in the detection results not matching the actual production requirements for both efficiency and quality. This makes it difficult to meet the actual needs of enterprises for efficient production scheduling and strict quality control, and restricts the smooth progress of subsequent production processes and the stable guarantee of product quality. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide an intelligent surface inspection method and system for medium and thick plates.

[0004] To solve the above technical problems, the technical solution of the present invention is as follows: A smart surface inspection method for medium-thick plates, comprising: Acquire image data of the steel plate to be inspected; The acquired steel plate image data is divided into a first detection area and a second detection area; A first detection precision and a second detection precision are preset in the detection model. The first detection precision is used to perform defect detection on the image data of the first detection area, and the second detection precision is used to perform defect detection on the second detection area.

[0005] As a preferred embodiment of the intelligent surface inspection method for medium and heavy plates described in this invention, the first inspection area includes the second head plate area, the second tail plate area, and the edge areas on both sides of the rolled steel plate, and the second inspection area includes the middle area of ​​the rolled steel plate.

[0006] As a preferred embodiment of the intelligent surface inspection method for medium and heavy plates described in this invention, the first inspection area further includes the head plate area, the tail plate area, and the edge areas on both sides of the flat-rolled steel plate, and the second inspection area further includes the middle area of ​​the flat-rolled steel plate.

[0007] As a preferred embodiment of the intelligent surface detection method for medium-thick plates described in this invention, the first detection area further includes all areas of the abnormal steel plate.

[0008] As a preferred embodiment of the intelligent surface inspection method for medium-thick plates described in this invention, after setting a first detection accuracy and a second detection accuracy of the pre-defined detection model, performing defect detection on the image data of the first detection area using the first detection accuracy, and performing defect detection on the second detection area using the second detection accuracy, the method further includes: Output the defect detection image data of the steel plate to be inspected, and mark the detected defects.

[0009] This invention also provides an intelligent surface inspection system for medium-thick plates, characterized in that it includes: The data acquisition module is used to acquire image data of the steel plate to be inspected; The region segmentation module is used to divide the acquired steel plate image data into a first detection region and a second detection region; The defect detection module is used to pre-set the first detection accuracy and the second detection accuracy of the detection model, and to perform defect detection on the image data of the first detection area using the first detection accuracy, and to perform defect detection on the second detection area using the second detection accuracy.

[0010] As a preferred embodiment of the intelligent surface inspection system for medium and thick plates described in this invention, it further includes a data output module for outputting defect detection image data of the steel plate to be inspected and marking the detected defects.

[0011] The beneficial effects of this invention are: (1) The present invention divides the detection area of ​​the steel plate and uses different detection precision to detect different detection areas, and enhances the identification precision for key areas with high defect incidence, thus ensuring the accuracy of defect identification.

[0012] (2) The present invention divides all areas of the abnormal steel plate into the first detection area, effectively reducing the false detection rate of abnormal steel billet. On the basis of improving the accuracy of defect identification and ensuring the surface quality of steel plate, the detection process is further optimized and the overall work efficiency is improved.

[0013] (3) Through reasonable algorithm design, the present invention realizes rapid identification and accurate recognition of the head and tail plates of steel plates, thereby improving the quality inspection efficiency of medium and heavy plate production and ensuring product quality. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1This is a schematic flowchart of the intelligent surface inspection method for medium-thick plates provided by the present invention. Figure 2 This is a schematic diagram illustrating the division of the steel plate detection area provided by the present invention. Detailed Implementation

[0016] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0017] Figure 1 The flowchart illustrates the intelligent surface detection method for medium-thick plates provided in this application embodiment. The method specifically includes the following steps: Step S101: Acquire image data of the steel plate to be inspected.

[0018] Specifically, an image acquisition device is fixedly installed in the steel plate inspection area. When the steel plate to be inspected passes through the inspection area, the image acquisition device can acquire images of the steel plate and transmit the acquired image data to the inspection model.

[0019] Step S102: Divide the acquired steel plate image data into a first detection area and a second detection area.

[0020] Specifically, the probability of defects occurring varies at different locations on a steel plate. Therefore, locations on the steel plate prone to defects are designated as the first inspection area, while locations less prone to defects are designated as the second inspection area.

[0021] See Figure 2 For coiled steel sheets, the first inspection area includes the second-head and second-tail areas of the coiled steel sheet, as well as the edge areas on both sides of the steel sheet; the second inspection area includes the middle area of ​​the coiled steel sheet. For flat-rolled steel sheets, the first inspection area includes the head and tail areas of the flat-rolled steel sheet, as well as the edge areas on both sides of the steel sheet; the second inspection area also includes the middle area of ​​the flat-rolled steel sheet. For abnormal steel sheets, all areas are included in the first inspection area.

[0022] Step S103: Pre-set the first detection accuracy and the second detection accuracy of the detection model, and use the first detection accuracy to perform defect detection on the image data of the first detection area, and use the second detection accuracy to perform defect detection on the second detection area.

[0023] Specifically, for the first detection area that is prone to defects, the control detection model is used to perform detection with the first detection accuracy; for the second detection area that is not prone to defects, the control detection model is used to perform detection with the second detection accuracy.

[0024] It should be noted that the accuracy of the first detection is higher than that of the second detection.

[0025] In this embodiment, the detection model used is the Pangu AI detection model.

[0026] Step S104: Output the defect detection image data of the steel plate to be inspected and mark the detected defects.

[0027] Specifically, the defects output by the detection model are visualized in the steel plate image, and an "updated" label is added to the image to distinguish between detected and undetected steel plates.

[0028] In addition, this application embodiment also provides an intelligent surface inspection system for medium and thick plates, the system including: a data acquisition module, a region division module, a defect detection module and a data output module.

[0029] Specifically, the data acquisition module is used to acquire image data of the steel plate to be inspected.

[0030] The region segmentation module is used to divide the acquired steel plate image data into a first detection region and a second detection region.

[0031] The defect detection module is used to pre-set the first detection accuracy and the second detection accuracy of the detection model, and to perform defect detection on the image data of the first detection area using the first detection accuracy, and to perform defect detection on the second detection area using the second detection accuracy.

[0032] It should be noted that the defect detection module includes the Pangu AI detection model for detecting defects in steel plates.

[0033] The data output module is used to output defect detection image data of the steel plate to be inspected and to mark the detected defects.

[0034] Therefore, the technical solution of this application divides the inspection area of ​​the steel plate, uses different inspection accuracies to inspect different inspection areas, and enhances the identification accuracy for key areas with high defect incidence, thus ensuring the accuracy of defect identification.

[0035] In addition to the above embodiments, the present invention may have other implementation methods; all technical solutions formed by equivalent substitution or equivalent transformation fall within the protection scope claimed by the present invention.

Claims

1. A smart surface inspection method for medium-thick plates, characterized in that: include: Acquire image data of the steel plate to be inspected; The acquired steel plate image data is divided into a first detection area and a second detection area; A first detection precision and a second detection precision are preset in the detection model. The first detection precision is used to perform defect detection on the image data of the first detection area, and the second detection precision is used to perform defect detection on the second detection area.

2. The intelligent surface inspection method for medium-thick plates according to claim 1, characterized in that: The first detection area includes the second head plate area, the second tail plate area, and the edge areas on both sides of the rolled steel plate, and the second detection area includes the middle area of ​​the rolled steel plate.

3. The intelligent surface inspection method for medium-thick plates according to claim 2, characterized in that: The first detection area also includes the head plate area, tail plate area and the edge areas on both sides of the flat-rolled steel plate, and the second detection area also includes the middle area of ​​the flat-rolled steel plate.

4. The intelligent surface inspection method for medium-thick plates according to claim 3, characterized in that: The first detection area also includes all areas of the abnormal steel plate.

5. The intelligent surface inspection method for medium-thick plates according to claim 1, characterized in that: After setting a first detection accuracy and a second detection accuracy for the pre-defined detection model, and performing defect detection on the image data of the first detection area using the first detection accuracy, and performing defect detection on the second detection area using the second detection accuracy, the method further includes: Output the defect detection image data of the steel plate to be inspected, and mark the detected defects.

6. An intelligent surface inspection system for medium-thick plates, characterized in that: include: The data acquisition module is used to acquire image data of the steel plate to be inspected; The region segmentation module is used to divide the acquired steel plate image data into a first detection region and a second detection region; The defect detection module is used to pre-set the first detection accuracy and the second detection accuracy of the detection model, and to perform defect detection on the image data of the first detection area using the first detection accuracy, and to perform defect detection on the second detection area using the second detection accuracy.

7. The intelligent surface inspection method for medium-thick plates according to claim 6, characterized in that: It also includes a data output module, which outputs defect detection image data of the steel plate to be inspected and marks the detected defects.

Images