HeightFusion_plus

We present a high-quality multi-modal roadside perception dataset named HeightFusion++. The sensing system is deployed at seven locations along a 600-meter continuous road segment and integrates multiple sensing modalities, eg. cameras, various LiDAR sensors, and 4D millimeter-wave radar. The dataset collects multi-modal data under diverse and challenging environmental conditions. Furthermore, a roadside-oriented multi-modal fusion method is proposed to enhance perception accuracy and robustness in complex scenarios, thereby providing a comprehensive data foundation for model training, algorithm development, and performance evaluation in roadside multi-modal perception.