Image feature extraction using a method derived from the Hough transform with extended Kalman filtering

The conventional implementation of the Hough Transform is inadequate in many cases due to its integrative effects of the discrete spaces. The design of an algorithm to extract optimal parameters of curves passing through image points requires a measure of statistical fitness. A strategy for image feature extraction called Tracking Hough Transform (THT) is presented that combines Extended Kalman Filtering with a Hough voting scheme that incorporates a formal noise model. The minimum mean-squares filtering process leads to high accuracy. Computing cost for real-time applications is addressed by introducing a converging sampling scheme. Extensive performance tests show that the algorithm can achieve faster speed, lower storage requirement and higher accuracy than the Standard Hough Transform.