Research on Multi-Scale Line Feature Extraction Algorithm Based on LSD

Line segment detector (LSD) algorithm is widely used in many systems that rely on image processing, and plays a decisive role in the stable operation of the system. To solve the problem that the line segment features extracted by traditional LSD algorithm are many short lines, and the line features cannot be extracted in some environments. In this paper, a multi-scale line feature extraction algorithm based on LSD is proposed. Firstly, a multi-scale space is added to the line segment extraction to reduce the missed detection of the line segment and combine the line segment with similar spatial position. Then input the image of the line segment extractor to sample and generate scale space to extract the preselected line segment set. Finally, the preselected line segment set is optimized by merging and fusion. Through experiment, this method is compared with the original method to extract the characters of more, at the same time to ensure the accuracy of the extracted features.