Research on Automatic Inspection of Factory Top Based on Drone Platform and Machine Vision
This study proposes a solution based on drones and machine vision for automatic inspection of accumulated water in the top manhole cover of a factory building.Due to the limited spatial positioning accuracy of drones,this article uses wide-angle photos and object detection algorithms to preliminarily locate the position of the manhole cover and reduce its offset in the image.Adopting real-time image retrieval and sharpness detection to solve the problem of image blurring caused by changes in wind speed.In order to address environmental factors such as seasonal changes,time differences,and weather conditions that affect image quality,a data augmentation strategy is adopted to improve the generalization performance of the model.Finally,a classification model was used to determine whether there were foreign objects around the manhole cover.Considering the diversity of foreign objects that caused the blockage of the manhole cover,a variety of training samples were collected to improve the model′s recognition ability.Through experimental verification,the method proposed in this paper exhibits high efficiency and accuracy in dealing with the above issues,and is of great value in improving the automation level of factory top inspection.
NETL
NSTL
万方数据
维普
