LinkingLines: Using the Hough Transform to Cluster Line Segments and Mesoscale Feature Extraction

Abstract

Linear feature analysis plays a fundamental role in geospatial applications, from detecting infrastructure networks to characterizing geological formations. In this paper, we introduce linkinglines , an open-source Python package tailored for the clustering and feature extraction of linear structures in geospatial data. Our package leverages the Hough Transform, commonly used in image processing, performs clustering of line segments in the Hough Space, and then provides unique feature extraction methods and visualization. linkinglines em-powers researchers, data scientists, and analysts across diverse domains to efficiently process, understand, and extract valuable insights from linear features, contributing to more informed decision-making and enhanced data-driven exploration. We have used linkinglines to map dike swarms with thousands of segments associated with Large Igneous Provinces in Kubo Hutchison et al. (2023).