Research On Binocular Ranging Technology Of Hazardous Chemicals Stacking Based On Improved ORB Algorithm

Maintaining the safe distance of "five distances"(stack distance,wall distance,top distance,column distance and channel distance)is the key to ensure the storage safety of hazardous chemicals.However,at present,the supervision of "five distances" mainly adopts manual inspection and lacks effective technical monitoring methods.Using binocular stereo vision ranging technology to carry out "five distance" monitoring of hazardous chemicals storage and stacking has important theoretical value and practical significance.Binocular stereo vision ranging is to use the pixel difference between the left and right camera imaging planes to obtain the three-dimensional coordinates of hazardous chemicals storage and the distance between stacks through image stereo matching.However,due to the similarity between the gray value of hazardous chemicals storage and the background wall and ground,and the similarity of stacking feature points is high,the binocular stereo vision ranging error is large,which can not meet the accuracy requirements of "five distance" monitoring.Aiming at the problem that the segmentation edge of hazardous chemicals warehouse stacking and warehouse background is not clear,and the accuracy of the extracted hazardous chemicals warehouse stacking boundary is not high,this paper proposes an improved OSTU algorithm,which uses genetic algorithm to find the best segmentation threshold of OSTU to segment hazardous chemicals warehouse stacking and background,so as to make the edges of stacking and background segmentation clearer and improve the accuracy of matching.Aiming at the problem of high similarity of feature points in hazardous chemicals storage and stacking images,an improved ORB algorithm is proposed.SURF algorithm is used to improve the feature detection of ORB algorithm.The feature points are given scale invariance by constructing three-layer image pyramid,and the feature points are described by rBRIEF descriptor.After rough matching,the wrong matching point pairs are deleted by feature registration based on consistency constraint,which greatly improves the accuracy of matching.Using the ORB algorithm,S-ORB(SURF-ORB)algorithm and the improved ORB algorithm,the matching ranging comparison experiment is carried out on the public data set and the hazardous chemicals storage and stacking image data set.The experimental results show that the average ranging error of the improved ORB algorithm is 0.35%,which is far less than the ranging error of the ORB algorithm and S-ORB algorithm.Combined with the improved OSTU algorithm and the improved ORB algorithm,a new OSTU-ORB ranging method is proposed.This method is used to locate the storage and stacking of dangerous chemicals.The average error of ranging is0.0758%,which is less than the improved ORB algorithm and meets the error range required by the storage and stacking of dangerous chemicals.The experimental results show that the improved OSTU algorithm solves the problem of high similarity between the gray value of hazardous chemical storage stacking and the background wall ground,and the improved ORB algorithm solves the problem of high similarity of stacking feature points.The dangerous chemical stacking ranging method of OSTU-ORB effectively improves the ranging accuracy of hazardous chemical storage stacking and meets the requirements of "five distances" for supervising hazardous chemical storage stacking.