Line Extraction-Split-Merge Algorithm

The MATLAB-based dichotomy method for line extraction with split-merge is a widely used image processing technique. This approach enables image segmentation with minimal distortion, facilitating more effective information extraction. The dichotomy-based line extraction method utilizes a binary search algorithm to isolate linear elements within images, followed by strategic segmentation and merging operations. This technique finds applications across numerous domains including computer vision, image processing, and machine learning. In implementation, the algorithm typically begins with edge detection using functions like edge() with Canny or Sobel operators. The core dichotomy process involves recursively dividing line segments at points of maximum curvature deviation, implemented through functions such as recursive_split() that calculate perpendicular distances between points and hypothetical lines. The merge phase employs distance and angle thresholds using regionprops() or custom functions to combine collinear segments. Within image processing, this dichotomy-based line extraction technology has been extensively applied to image segmentation and object detection tasks. Key MATLAB functions involved include poly2mask() for conversion between polynomial representations and binary masks, and houghlines() for alternative line validation. The research and application of MATLAB's dichotomy method for line extraction with split-merge capabilities holds significant importance for advancing automated image analysis systems.