Simulation of Extracting Complex Texture Features of Cracks in Old Single Landscape Buildings
Old single landscape buildings usually have complex textures,which are affected by changes in lighting conditions and the shadows of the building itself,resulting in increased difficulty and lower accuracy in extracting complex texture features of surface cracks.Therefore,a simulation was proposed to extract complex texture features of cracks in old single landscape buildings.Firstly,the mask was constructed.And then,noise was eliminated by fast Fourier transform algorithm.Secondly,histogram of old single landscape building images was counted,and the gray level was corrected to obtain clearer images.After the pseudo-cracks with gradient were eliminated,the filtered spatial image could be obtained.Moreover,a Sobel Operator(Sobel)template was constructed by the edge detection.Meanwhile,the response value of edge detection was calculated.Finally,all pixel points were integrated into a feature set for calculating the target information loss of all cracks.Thus,the best feature extraction for texture cracks was a-chieved.The experiment results show that the proposed method has better preprocessing effect on complex texture im-age,higher extraction accuracy,and can even extract the smallest cracks,which is helpful for technicians to repair and ensure building safety.
Old single landscape architectureComplex texture of buildingCrack feature extractionGaussian fil-terSmooth noise reduction
万方数据
维普
