Research And Design Of Surface Defect Inspection Of Ring Cooling Tray Based On Machine Vision

Cracks and abnormal wear defects on the wheel treads of the trolley during trolley operation.These two types of defects have the uncertainties of quantity,shape,and position.At present,the inspection of trolley wheel tread defects mainly relies on manual inspection.However,this kind of detection method has low efficiency,low accuracy,and is easy to miss detection,which is difficult to meet the requirements of industrial intelligence and smart manufacturing.In order to solve the problems of the above detection,This article will based on machine vision inspection technology,research on-line trolley wheel tread defect detection system to realize the detection and classification of trolley wheel tread defects.The main research contents of this paper are as follows:1.According to the requirements of trolley wheel tread defect detection and actual working conditions,choose the different kings of industrial cameras,optional lens and fill lights.For the display of the image detection results of the trolley wheels,a static historical monitoring image area and a real-time monitoring image area are designed,and there are related detection data,the purpose is to provide data support for subsequent image recognition and classification.2.Image processing and defect detection.First,the image is preprocessed,the purpose is to eliminate image noise and enhance the contrast of image contour information and then,perform image segmentation.At last,the tread area and potential defect area of the platform wheel were extracted.Image segmentation of trolley wheels using an iterative algorithm,extraction of trolley wheel tread area based on morphology and Hough circle transform,using an improved HOG feature extraction algorithm based on information entropy weighting to extract potential defects.The selection and improvement of the image defect area algorithm is based on whether it can effectively extract the area features and reduce the amount of calculation.The selection and improvement of the image defect area algorithm is based on whether it can effectively extract the area features and reduce the amount of calculation.There are two types of defects in the tread of trolley wheels: cracks and abnormal wear.However,potential defect areas include not only cracks and abnormal wear defect areas,but also pseudo defect areas.To construct the feature description of the potential defect area based on the defect feature,first of all,the geometric feature,texture feature and location feature are used for the training of the random forest algorithm model;then the real defect detection is performed on the potential defect area to identify the defect type.3.Design defect detection system and experimental analysis.Design the overall scheme of detection system according to the requirements of defect detection technical index,and verify the accuracy and effectiveness of the algorithm.The experimental results show that the system can obtain the surface image of the trolley wheel that meets the inspection requirements,defect detection accuracy is higher than 97%,The detection accuracy of the defect type reaches more than 90%.Therefore,the detection method proposed in this paper can effectively solve the detection problem of the surface defect of the trolley wheel and detection accuracy meets industrial production requirements.