Building Edge Detection Based on Dyadic Wavelet (TA-P8)

Author(s) : Qiming Qin (Institute of Rs&GIS, Peking University, China) Wenjun Wang (Department of Communication Engineering, Yanshan University, China) Sijin Chen (Institute of Rs&GIS, Peking University, China)

Abstract : Abstract: In the high resolution satellite image, edge detection often lead to lose the useful detail information from satellite image or bring about some spurious edge. To solve these problems, the dyadic wavelet with the multi-resolution analyzing is used to detect the buildings edge from satellite. Via the dyadic wavelet decomposing, the building satellite image of a certain appropriate scale can be obtained, and figured out the edge data of the horizontal and the upright directions respectively, then the grads vector module of the buildings features was calculated and the points in the edges of an image were gotten, at last these edge points of the buildings were checked and traced, the checked edge of buildings were obtained. An experiment of edge detection presents the dyadic wavelet approach can depress the effect of noise and examine exactly the edge data of the building object by rule and line.

Menu