【摘要】 论文为了降低复杂场景中基于单一传感器进行目标检测的局限性，提出了一种特征级的毫米波雷达和图像融合的目标检测方法（SPCRF-Net）。该方法将毫米波雷达原始数据预处理成固定大小的线段并映射到图像中，引入金字塔池化处理毫米波雷达数据；对图像采用VGG16作为主干网络进行特征提取，并在每一层中融合毫米波雷达特征和图像特征。在融合层次中引入SE注意力模块增强高级别特征感知能力，并构建了一种融合结构（PFPN）强化特征提取。实验表明该方法有效减少了目标的漏检情况，提升了模型目标检测的性能。更多还原

【Abstract】 In order to alleviate the limitations of object detection based on a single sensor in complex scenes,the paper proposes an object detection method(SPCRF-Net)based on millimeter wave radar and image at feature-level fusion. This method preprocesses radar raw data into fixed-size line segments and maps them to the image,and introduces pyramid pooling to process radar data. And then this method uses VGG16 as the backbone for image feature extraction,and integrates radar features and image features in each layer. In the fusion level,the SE attention module is introduced to enhance the ability of high-level feature perception,and a fusion structure(PFPN)is constructed to enhance feature extraction. Experiments show that this method can effectively reduce the missed detection of targets and can improve the performance of object detection.更多还原