Object Recognition by Fourier Descriptor

Fourier Descriptors(FD) is a common way for representing the boundary of an object. In this paper, an algorithm has been implemented to do object recognition by using FD. This is applied to various tool object, and is tested. This implementation contains two parts: image acquisition and object recognition. Appropriate lighting, viewing angle, and strong contrast of background and object are taken into account in this aspect. Minimum distances are calculated by using FD's and boundary matching among objects on the process of object recognition. Rotation, translation and scaling of the object will not influence the performance of the algorithm. Experiments show that we can use only one fourth of 1024 FD coefficients to do raped object recognition.