Detection of Line Objects by Means of Gabor Wavelets and Hough Transform

Archives of civil engineering

article

Urszula Marmol¹  Natalia Borowiec¹ 

[1] AGH University of Science and Technology, Faculty of Mining Surveying and Environmental Engineering

关键词: edge detection;    line objects;    Gabor wavelets;    Hough transform;   

DOI  :  10.24425/ace.2020.134401

学科分类：建筑学

来源: Polska Akademia Nauk * Instytut Podstawowych Problemow Techniki / Polish Academy of Sciences, Institute of Fundamental Technological Research

【 摘 要 】

This article presents a method for detecting linear objects with a defined direction based on image and lidar data. It was decided to use Gabor waves for this purpose. The Gabor wavelet is a sinusoid modulated by the Gauss function. The orientation angle of the sinusoid means that the waveform can only operate in strictly defined directions. It should, therefore, provide an appropriate solution to the problem posed by the publication. The research problem focused in the first stage on determining the approximate location of only the analysed objects, and in the next step on correct and accurate detection. The first stage was carried out using Gabor filters, the second - using the Hough transform. The tests were performed for both laser data and image data. In both cases, good results were obtained for both stages: approximate location and precise detection.

【 授权许可】

Unknown