Research On Safety Distance Detection Technology Of Stacking In Hazardous Chemical Warehouse Based On Binocular Vision

The safety distance of hazardous chemicals stacking is an important index to ensure the safety of storage,but the relevant technical means for the detection of the safety distance is not mature enough,and it is difficult for the safety department to supervise it in real time.Using binocular distance measurement can monitor the safety distance in real time,which is of great significance to the security of hazardous chemical warehouse.However,due to the complex environment of the chemical warehouse,the feature points of the stacking image are similar and the stacking texture is sparse,which leads to the difficulty in matching the binocular vision and affects the ranging accuracy of the stacking.In this paper,AKAZE-GMS binocular matching algorithm is proposed to detect the safe distance of stacking by using nonlinear multiscale feature detection and motion smoothing constraint to eliminate mismatching.In this algorithm,nonlinear diffusion filtering is introduced to construct nonlinear scale space to extract the feature points of the stacking,which preserves the details and edges of the stacking and improves the repeatability and uniqueness of the feature points.M-LDB descriptor is used to describe the feature points of the stacking,and hamming distance is used as similarity measure to perform rough matching of the feature points.Then,the grid-based motion statistics(GMS)method was used to further purify and eliminate the false matching point pairs.The algorithm solves the problem that the feature points of the stacking images are similar and the texture is sparse,which makes the matching accuracy of the stacking images reach 96%,which is about 24% higher than the classical SIFT algorithm.The minimum average error of the distance detection results between the stacking images is kept within 0.5%.In the experimental process,it was found that the complex environment of hazardous chemicals and uneven illumination would affect the matching effect of AKAZE-GMS algorithm.To solve this problem,a piecewise linear transformation based method was further proposed to enhance the stacking image.The experimental results show that the AKAZE-GMS algorithm greatly improves the matching logarithm of feature points,and the average matching rate and accuracy are high,and the AKAZE-GMS binocular matching algorithm performs well in the actual detection of hazardous chemical stacking safety distance.The AKAZE-GMS binocular matching algorithm proposed in this paper can effectively solve the problem of difficult three-dimensional matching for the stacking of hazardous chemical warehouses,and the measurement accuracy of the stacking distance is high,which is suitable for the supervision of the safety distance of the stacking of hazardous chemical warehouses.