Road Triangle Detection for Non-Road Area Elimination Using Lane Detection and Image Multiplication

Abstract

The background has become the key issue in maintaining the accuracy of final analysis for object detection in the development of an image processing algorithm. Therefore, this paper focuses on intelligent transport system (ITS), in which some of the background characteristics such as trees, road divider, and buildings interfere in the detection system algorithm. Therefore, this paper presents an algorithm that can remove the unwanted background, outside the road area boundaries for dynamic video footage. Using the onboard camera to capture the road traffic, the background is always moving in motion together with the foreground; therefore, a region of interest that focuses only on the road region needs to be established. The algorithm consists of three main components: lane detection, vanishing point and image multiplication. From the three components, other methods are applied, namely Hough transform, line intersection, image masking and image multiplication, which are combined together to create the background subtraction system. In the final analysis, the test results under various road conditions show a good detection rate and background removal.