Vehicle Distance Estimation from a Monocular Camera for Advanced Driver Assistance Systems

The purpose of this study is to propose a framework for accurate and efficient vehicle distance estimation from monocular camera. proposed consists transformer-based object detector, depth estimator, predictor. detector detects various objects that are mostly symmetrical an image captured by the camera provides type each coordinate information bounding box around object. estimator generates map image. Then, boxes overlapped with extract features object, such as mean depth, minimum maximum present then trained three models—eXtreme Gradient Boosting, Random Forest, Long Short-Term Memory—to predict actual between based on (including its coordinates size), extracted features. proposes including trimmed excluding background pixels but within evaluation results show outperformed existing studies.