Accurate Association of Radar-Vision Measurements Based on Weighted Euclidean Distance

Associating radar measurements with vision measurements is essential for radar-vision fusion. Most radar-vision association methods rely on Euclidean distance. However, these methods have overlooked measurement errors, leading to a significant amount of erroneous associations. To address the limitations of existing methods, we propose an association method based on weighted Euclidean distance between radar measurements and vision measurements. This paper analyzes the measurement errors of radar and vision and uses this analysis as the basis for calculating weighted Euclidean distance, to make weighted Euclidean distance sensitive to measurement errors and reduce the occurrence of association mistakes caused by such inaccuracies. The obtained weighted Euclidean distance is then used as a cost matrix to perform association using the Hungarian algorithm. The proposed method is validated through simulations, and the experiments demonstrate that our approach achieves higher accuracy.