Given the continuous advancements in transportation technology and the rising number of vehicles, coupled with the pressing need to address frequent traffic accidents, there has been a surge of interest and research in various domains such as intelligent road testing, traffic monitoring, automatic driving, and assisted driving technologies. To address the limitation of the multi-resolution sensor fusion (HRFuser) model in adequately capturing essential features during the feature extraction process and detection, we propose an Enhanced HRFuser (EHRFuser) model. This model integrates an Efficient Channel and Spatial Attention-Bottleneck (ECSA-Bott) module and an Enhanced Cascade Region-based Convolutional Network (ECascade R-CNN). The ECSA-Bott module incorporates multiple attention mechanisms to focus on pertinent information across channel and spatial dimensions during convolutional feature extraction. We construct the ECascade R-CNN by integrating the ECSA module into the detector, thus enhancing key details and improving the information representation of the fused feature map. Comparative experiments demonstrate superior performance of the EHRFuser over the HRFuser across all metrics, with a notable 1.7% improvement in the AP0.5 indicator, validating its enhanced performance in target detection tasks.