Adaptive linear feature detection based on beamlet

Linear feature detection is very important in computer vision, image segmentation and pattern recognition. Traditional linear feature detectors based on pixel processing each by each may fail to detect out lines in image with low SNR. A fast discrete beamlet transform and an adaptive method of linear feature detection are proposed, which can detect lines with any orientation, location and length. The scale parameter can be adaptively determined by histogram of beamlet energy function distribution. Experiment results prove the efficiency of the method proposed even in image with very low SNR.