An Improved Canny Edge Detector and its Realization on FPGA
Problems with traditional algorithm:
In the traditional Canny detector, the threshold used in the algorithm needs to be set beforehand according to the background-to-image ratio of the image to be processed. However, since the edge-to-ground ratio of an image usually can not be known in advance, and it varies quite a lot with the changes of the illumination conditions, the traditional Canny algorithm can not always give good edge detection results. Therefore, it is necessary to find a good way so tha the threshold in the Canny algorithm can be obtained automatically instead of manually for a given image under the given illumination conditions.
Self-adapt threshold Canny algorithm
A new self-adapt threshold Canny algorithm is proposed in this paper to solve the first problem. A pipelined implementation on FPGA for this new algorithm is also designed to solve the second problem. Experiment results are also given to show the efficiency of the proposed method. In this paper, a selfadapt threshold Canny edge detection algorithm is proposed. In this new Canny edge detector, instead of being set by manual, the threshold can be set automatically by the algorithm itself. Therefore, this new algorithm is more adaptable to the changes of environments and illumination conditions. The effectiveness of the proposed method has been shown by experiments. To meet the needs of real time processing, they have designed a pipelined implementation on FPGA for this improved edge detection algorithm. Compared with the implementation in a PC based system, this pipelined implementation on FPGA takes much less implementation time and can therefore be used for the mobile robot vision system which is very strict for the real-time performance of its vision system.
Wenhao He and Kui Yuan, “An improved Canny edge detector and its realization on FPGA,” 2008 7th World Congress on Intelligent Control and Automation, 2008, doi: 10.1109/wcica.2008.4594570.