Edge Detection from High-resolution Remotely Sensed Imagery Based on Gabor Filter in Frequency Domain

Edge detection is important to image processing, and there are many approaches to detect the edge features from image. However, the high-resolution remotely sensed imagery including complex features is difficult to extract the edge features. This paper introduces a new method to detect edge from the high-resolution remotely sensed imagery. By using FFT, the high-resolution remotely sensed imagery is transformed into the magnitude spectrum image (frequency image). Then, the frequency image is analyzed by the Radius sampling and angle sample. Finally, Gabor filter is used to extract the edge features by setting the proper parameter. The experimental result shows that the proposed method gives a better solution for extract edge feature from the high-resolution remotely sensed imagery.