Research on Visual Fault Detection and Location of Ventilation Machinery Instrument Panel in Complex Background Environment
Ventilation mechanical instrument panels often have the image features of inconsistent color,brightness,and texture changes due to shadows or partial occlusion in complex background environments,resulting in a contrast decrease between the faulty area and the surrounding environment,so it is difficult for manual methods to accurately locate the faulty area.To address these is-sues,a visual fault detection and localization method for ventilation machinery instrument panels is designed.Kinect cameras are used to extract ventilation mechanical instrument images and the histogram equalization is performed to adjust the brightness and hue of the image,enhancing the local contrast between the fault contour and the background.By using an improved pixel correlation segmenta-tion algorithm to segment the image features,the dashboard area in the image is extracted from complex backgrounds.Deep convolu-tion networks are used to detect the fault contour on segmented instrument panel images.the centroid coordinates of the positioning target(fault contour)is calculated,the centroid position is taken as the target point,and it is mapped to the constructed projection imaging spatial coordinate system,which achieves the high-precision positioning of displaying the fault area on the dashboard.The ex-perimental results show that after applying this method,the contrast distinction has a significant enhancement between the fault area and the surrounding environment,with a high detection and positioning accuracy.
万方数据
维普
