Convolutional neural network-based intelligent identification method for medium-thickness plate indentations
A convolutional neural network, intelligent identification technology, applied in the field of intelligent inspection and identification of medium and heavy plate indentation, can solve the problems of uneven internal texture, unevenness, low identification rate, etc., to reduce batch quality accidents and meet surface quality standards , the effect of facilitating processing
Inactive Publication Date: 2019-07-19
福建三钢闽光股份有限公司 +1
View PDF3 Cites 4 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
The research shows that indentation defects are characterized by boundaries, uneven internal textures, and unevenness. According to the boundary of indentation defects, it is difficult to detect and identify them accurately by using traditional indentation recognition technology. relatively low
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreExamples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0030] A method for intelligent identification of plate indentation based on convolutional neural network, said identification method comprising the following specific steps:
[0031] Step 1: Acquisition of image data on the surface of the medium-thick plate: using image acquisition equipment to collect image data on the surface of the medium-thick plate passing through the production line;
[0032] Step 2: Identification of indentation defects on the surface of medium and thick plates: This process includes the following specific steps:
[0033] a) Preprocessing the image data collected in step 1 by image filtering;
[0034] b) Use the sobel operator to perform edge calculation and detection, and calculate the gradient value of each pixel through the horizontal and vertical operators to obtain its gradient map;
[0035] c) Carrying out weight overlapping processing with gradients in the X and Y directions, respectively, to obtain gradient images with enhanced gradients;
[...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a convolutional neural network-based intelligent identification method for medium-thickness plate indentations. The method comprises the following specific steps of acquiring the image data of the surface of a medium-thickness plate, identifying indentation defects on the surface of the medium-thickness plate, and intelligently classifying and identifying the identified indentation defect images. The method is based on a defect identification method of image filtering, sobel edge detection, gradient enhancement and threshold value automatic adjustment processing, and then the intelligent classification recognition is carried out on the indentation defect images by combining a convolutional neural training model, so that the indentation defect detection rate in an actual medium plate production line reaches 96.83%, and the method has the obvious advantage than a traditional identification method of which the detection rate is 27.55%.
Description
technical field [0001] The invention relates to the field of intelligent inspection and identification of indentation of medium and thick plates, more specifically, an intelligent identification method for indentation of medium and thick plates based on convolutional neural network. Background technique [0002] In the production of medium and heavy plates, it often occurs that the steel plate is changed or rejected due to indentation. Due to the small shape of indentation defects, the recognition effect of traditional detection methods is poor, and it is difficult to detect accurately, and the indentations basically show the characteristics of batches, which not only affects the indicators and economic benefits of the product, but also seriously damages the image of the product. [0003] Traditional indentation recognition methods only use image filtering and edge detection to identify defects. The research shows that indentation defects are characterized by boundaries, un...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06T7/00G06T7/13G06T7/155
CPCG06T7/0004G06T7/13G06T7/155G06T2207/20084
Inventor 陈玉叶詹光曹郑芳垣陈旸
Owner 福建三钢闽光股份有限公司