A test framework for the accuracy of line detection by Hough Transforms

The detection of lines in an image is an important task. In the fact that many works have been done on line extraction, there is a lack of a comprehensive comparison of the so far proposed algorithms. The design and implementation of a framework to test line detection algorithms on intensity images is described in this paper. Our test framework is applied to compare the correctness and precision of lines extracted by Standard Hough Transform, Progressive Probabilistic Hough Transform and a proposed method based on Standard Hough Transform. The correctness of the extracted lines relates to global accuracy whereas the precision concerns accuracy at a local level. The well-known Standard Hough Transform (SHT) and Progressive Probabilistic Hough Transform (PPHT) are two of the most efficient algorithms for line detection. SHT can detect almost straight lines in the image, and it is highly resistant to noise. Line segments are effectively found by PPHT. However, this algorithm has lower accuracy than SHT. The proposed method based on SHT overcomes this. It contains three extensions: the technique of accumulation, the application of a local maxima rule, and the detection of line segments. The test framework enables us to evaluate the advantages and disadvantages of the three Hough Transform algorithms by analyzing the results of line extraction.