A Fourier Descriptor for Bone Point Segmentation using inner distance in remote sensing images

In most cases, especially in remote sensing targets, the contour of the object is used to describe the features of the object, so the shape descriptor plays an indispensable role in the target detection and recognition. In recent years, more and more shape descriptors have come to the fore, but many descriptors ignore the details of the shape, such as the traditional shape descriptor of centroid and contour descriptor (CCD) and shape context descriptor (SC). The inner distance of a shape has strong robustness for describing shape features, and the inner distance shape context9(IDSC) is a good example. Therefore, this paper proposes a method of using the inner distance to find shape bones and segment the shape contour by these bones, finally performing Fourier transform to form shape feature (FD-IDBS). In matching time, we using the distance difference to perform shape matching. What is commendable is that its discriminability and robustness is strong, process is simple and matching speed is fast. More importantly, the experiment results show that the shape descriptor has higher retrieval accuracy.