FPGA implementation of histogram-based thresholding

Edge detection or segmentation results in a binary image. Thresholding is most often used for this purpose. The final binary image depends significantly on used threshold value. Several methods have been published to solve the problem of determining the optimal threshold. In this paper we propose FPGA implementation of Otsu's method for automatic unsupervised and nonparametric determination of picture threshold value based on picture histogram. This method has not been used in FPGA circuits so far because of relatively complicated implementation of the calculation algorithm.