Approach to linear feature extraction for infrared image

In this paper, an approach to linear feature extraction for infrared image is presented, which consists of three major modules: image preprocessing by using fuzzy feature transformation and fuzzy enhancement, edge strength map and direction map generation by using low pass filter and multi-dimension edge detector and linear feature extraction by using gradient profile maximum method. Comparison with some other edge detection methods, accuracy and robust experiments are done to testify its better position accuracy and reliability. Computational simulation and experiment results show that the proposed algorithm can solve the linear feature extraction problem for infrared image.