Line Detection in Image Based on Edge Enhancement

The Radon transform is an important tool to detect line features in the image and it has strong anti-noise ability. The traditional method, which the first step is to do edge detection and then to do the Radon transform, exists some defects. Therefore, this paper has made some improvements, that is, using the empirical mode decomposition (EMD) method to decompose the image under different scales. The intrinsic mode functions (IMFS) are obtained. The first IMF, on finest scale, is superposed with the gradient of the image, so as to enhance the edge of the image. This method can avoid the interference information that directly doing edge detection brings. In addition, the global Radon transform can not detect the length of the segment as well as the endpoint information, and also has difficulty in detecting the short segments. For these defects, the idea of Local Radon transform has been adopted in this paper. The main method is to move a sliding window on the image, then do Radon transform on the region that covered by the sliding window. The experiment results show that our method is efficient to detect the line features in image.