Crack Detection via Differential Excitation and Gradient Orientation Based on Descriptor

Crack detection and measurement in civil structures has been a constant field of research. In this paper, differential excitation, gradient orientation which is the important components of descriptor is proposed which is a new approach in the field of crack detection. Differential excitation component is defined as a function of two components. i. e. , the relative intensity differences of current pixel with its neighbors and the intensity of current pixel whereas the orientation component is the gradient orientation of current pixel. It is done by introducing local spatial information, where an image is divided into number of blocks and histogram were computed for each blocks and concatenated. Based on this crack detection and further measurement of crack is done. Experimental result shows that this technique exhibits better performance and crack detection is done more efficiently.