Straight line extraction via multi-scale Hough transform based on pre-storage weight matrix

摘要

Considering the low efficiency and accuracy of existing straight line extraction methods from large-scale imagery, a multi-scale Hough transform (HT) method based on the pre-storage weight matrix is proposed. Improvements of the proposed HT include: using the pre-storage weight matrix to save storage space and improve the efficiency of the HT, applying a multi-scale method to enable the detection of smaller features from large-size images, using overlapping adjacent tiles to avoid the fragmented lines, optimizing the dimensions of the discrete transform domain, normalizing the accumulator to extract off-centred lines and using a dual threshold for distributed line detection. The experimental results show that the proposed algorithm is not only more efficient and robust but also can obtain richer and more accurate information about features, especially for large-size images.