Steel plate surface defect detection method based on yolov3-spp network model
A network model, defect detection technology, applied in character and pattern recognition, image analysis, image enhancement and other directions, can solve problems such as low adaptability and accuracy, and achieve the effect of improving accuracy and expanding sample size
Pending Publication Date: 2021-07-23
SOUTH CHINA UNIV OF TECH
View PDF4 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0003] Traditional steel plate surface detection technologies mainly include manual detection methods, detection methods based on electromagnetic waves or ultrasound, pattern recognition or machine learning methods, with low adaptability and accuracy
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0031] The present invention will be further described below in conjunction with specific examples.
[0032] Such as Figure 1 to Figure 4 As shown, the steel plate surface defect detection method based on the yolov3-spp network model provided in this embodiment uses the yolov3-spp network model (composed of backbone, spp module, neck and detection head) to train and analyze the steel plate surface defect pictures Test, which includes the following steps:
[0033] 1) Obtain the steel plate surface defect picture, be divided into training set and test set, comprise the following steps:
[0034] 1.1) Obtain pictures of various types of steel plate surface defects including inclusions, scratches, rolling scale, cracks, pitting surface or plaque defects, 300 pictures of each type of defect, a total of 1800 pictures, and the size of the pictures is 200px* 200px;
[0035] 1.2) Extract 80% of the steel plate surface defect pictures, that is, 1440 as a training set, and the remaini...
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 steel plate surface defect detection method based on a yolov3-spp network model, which comprises the following steps: 1) acquiring a steel plate surface defect picture, and dividing the picture into a training set and a test set; 2) inputting the pictures of the training set into a yolov3-spp network model, performing iterative training, adjusting weight parameters of the network model, and obtaining a weight file of the network model after a set number of iterations is reached; and 3) loading the weight file to the yolov3-spp network model, and predicting the pictures of the test set to obtain a detection result. According to the method, the spp module in the yolov3-spp network model is taken as an effective means for realizing fusion of local features and global features, the structural information of the defect in the picture can be fully extracted, and through learning of the training set, the accuracy rate of steel plate surface defect detection on the test set can be effectively improved.
Description
technical field [0001] The invention relates to the technical field of steel plate defect detection, in particular to a steel plate surface defect detection method based on a yolov3-spp network model. Background technique [0002] Steel plate is an essential basic material in the machinery industry, and is widely used in automobiles, home appliances, electric power and other industries. Affected by factors such as equipment and production process limitations, various defects will inevitably occur on the surface of steel plates, such as cracks, inclusions, plaques, pitting surfaces, rolling scales, scratches, etc. In order to improve the production of steel plates To ensure the quality of the steel plate, attention must be paid to the detection of surface defects of the steel plate. [0003] Traditional steel plate surface detection technologies mainly include manual detection methods, detection methods based on electromagnetic waves or ultrasound, pattern recognition or mac...
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 IPC IPC(8): G06T7/00G06K9/62G06K9/46
CPCG06T7/0004G06T2207/20081G06T2207/20084G06T2207/30136G06V10/454G06F18/253
Inventor 许玉格杨舒乔吴宗泽邓木清
Owner SOUTH CHINA UNIV OF TECH