This paper proposes a window-based gray-scale inverse Hough transform. The proposed algorithm can be applied to any gray-scale image and allows line extraction according to filtering conditions. The technique reduces the computational time and the memory storage requirements of the inversion procedure considering the image as a set of sub-windows and working only with the necessary gray-scale values. Also, a new improved definition of the representation of a straight line in the accumulator array is used that takes into account the image quantization effects and allows the extraction of lines with variable thickness. The pixels of the resulting lines in the final image appear in their exact position and have gray-scale values as in the original gray-scale image.
展开▼