Research On Crack Mask Optimization Algorithm And Crack Measurement Based On Ant Colony Algorithm

The structure is in disrepair for years,the design and construction are not standardized,and the structure caused by severe natural disasters has potential safety hazards.Therefore,the structural safety inspection is particularly important.The first signal for judging structural problems is the occurrence of cracks in its structural components.How to analyze and study cracks quickly and accurately is a direction that engineers and technicians have been tirelessly exploring.The main research work of this paper is as follows:(1)This article proposes a crack detection system,introduces the main building blocks of the system and the design process,using this as the skeleton to describe each module in the system in detail through each chapter,and elaborate the operating logic and rules of the detection system.(2)By collecting crack photos of different scenes in reality,labeling the image of the crack boundary through Label me software,using the labeled image as the training set of the deep learning algorithm for training,after the training is completed,the experimental image is passed through Mask-RCNN to identify the crack,Get the mask information image of the crack.Extract the crack mask information through python,read the coordinates of the mask,the pixel value of the crack picture,use Open CV-python to grayscale the picture and extract the gray value of the crack picture,and calculate the gray level of the image through the canny operator Derive the image gray gradient value,use Open CV to extract the image gray gradient value,complete the mask information extraction and processing.(3)Optimize the crack mask through the ant colony algorithm,the heuristic guide information is the gray value of the image,the heuristic guide function obtained through the above processing is set to the gray gradient value of the image,and the initial pheromone is set as a constant,Iterate through the release of pheromone by K ants,and record the iterative information coordinates.When the iterative information coordinates are unchanged,the optimal solution is obtained,which is the optimized crack information.(4)Calculate the length and width of the crack,extract the coordinate information of the crack contour,obtain the crack width pixel and the crack center line through the algorithm of the outer expansion circle,obtain the crack length pixel,and obtain the crack length and width through the conversion of the same proportional distance of the pixels value.(5)Carry out the actual demonstration of the algorithm,compare and analyze the processing results of the algorithm in different scenarios,carry out the ant colony algorithm optimization clarity and accuracy comparison experiment,and the accuracy comparison experiment of different crack parameters to verify the universal applicability of the algorithm under different conditions.