1, the input image
2, histogram equalization, so that images are normalized to the magnitude 0-1
3, to reduce the influence SAR image speckle by a mean filter
4, two threshold values
5, respectively, in the target region and the hatched region select the maximum communication area, communication if the target area is larger than the maximum given threshold value, a detection target is considered complete image preprocessing
6, the smallest circumscribed rectangle solved, particularly around the target step to draw a rectangle side length l1Xl2 ensure tangent rectangle object. Thereafter, each time rotating the rectangle at an angle a, up to meet certain criteria. There are guidelines for the determination of the target - background area than the maximum, minimum and perimeter boundary and so on up to the intersection points. It is generally accepted that the border point up to the highest accuracy criteria.
7. The four boundary points of the boundary calculated above, to calculate the azimuth of the target