Journal Search Engine
The Journal of The Korea Institute of Intelligent Transport Systems Vol.24 No.6 pp.275-293
자율주행차 도로파손 탐지를 위한 시각인지 기반 특징분석 및 이미지 필터 연구
Feature Analysis and Image Filter-Based Enhancement of Visual Perception for Road Damage Detection in Autonomous Vehicles
Abstract
Autonomous vehicles rely on camera-based visual perception systems to make driving decisions, yet low-light and low-contrast conditions often lead to object detection errors. This study investigates the causes of visual misperception in detecting road surface damage and proposes an image preprocessing filter to mitigate these issues. Pixel brightness data were used to classify images into correctly detected and undetected sets, followed by histogram-based feature analysis across eight indices to identify key factors contributing to detection failures. Based on these findings, a brightness-adjustment preprocessing filter was developed and evaluated. Experiments show that the proposed filter reduces the miss-detection rate by about 50% compared to the baseline. These results demonstrate that pixel brightness-based preprocessing can significantly improve autonomous vehicle perception of road damage under adverse lighting conditions.
