| Maintaining the distance between pallets,walls,columns,lights,and beams of hazardous chemical storage within the specified safe distance range is a key factor in ensuring the safety of goods storage.The use of binocular stereo vision technology can detect and monitor the safe distance of storage stacks.However,due to poor lighting conditions and large variations in the warehouse,as well as similar shapes between goods and small contrast with the background,the use of binocular vision technology directly matches the entire image in a three-dimensional manner,resulting in low matching accuracy and slow algorithm speed,which does not meet the requirements of real-time monitoring.This article adopts object detection method and improved binocular stereo matching algorithm to effectively achieve safe distance detection for hazardous chemical storage stacking.The main research content includes:(1)To overcome the influence of background on stereo matching,a variational automatic encoder(VAE)was used to expand the dataset,and YOLOv8 s object detection algorithm was used to identify the stacking area.Then,image recognition and stereo matching were performed,resulting in a success rate of 99.6% for stacking detection and a recognition speed of 21.5 fps/s.The stacking area can be extracted in real-time,filtering out useless background information.(2)An improved Shi-Tomasi algorithm was proposed.The algorithm extracts corresponding feature points at different scales by constructing a Gaussian difference image pyramid,overcoming issues such as lighting,noise,and feature similarity in hazardous chemical stacking matching.At the same time,adaptive local affine validation is used to filter out incorrect matching point pairs,further improving the stereo matching effect of stacking images.(3)This algorithm has been compared with ORB algorithm and Shi-Tomasi-ORB algorithm in experiments.The results show that the matching accuracy of this algorithm remains around 97%,with an improvement of about 30%.The average error of distance measurement between stacks is within 1%,which is superior to other algorithms.This article effectively solves the problems of low matching accuracy and large ranging error caused by the influence of lighting and background interference in the binocular ranging of hazardous chemical warehouse stacks through the identification of stacking areas and the introduction of an improved Shi-Tomasi algorithm in ORB.This method meets the requirements for safe distance between hazardous chemical warehouse stacks and provides strong technical support for the storage safety of stacks. |
