【摘要】 随着无人车等智能系统的发展，对三维目标识别技术的需求日益增长。针对传统三维目标识别方法在实时性和数据量方面的局限性，提出了一种基于仿人眼视网膜特性的三维目标识别方法。通过建立仿人眼视网膜映射模型，实现了对目标的空间变分辨率成像，有效压缩了非感兴趣区域的数据，提高了对感兴趣目标的识别效率。最终在公开数据集KITTI与Jetson Xavier Nx开发板通过实验验证，证明了所提方法的有效性。更多还原

【Abstract】 With the development of intelligent systems such as unmanned vehicles, there is a growing demand for 3D target recognition technology. In this work, for the limitations of traditional 3D target recognition methods in terms of real-time and data volume, a 3D target recognition method based on the bio-inspired retina-like characteristics was proposed. By establishing a bio-inspired retina-like mapping model, spatially variant resolution imaging of the target was realized, which could effectively compress the redundant information in the non-interested region and improve the recognition efficiency of the interested target. Finally,the effectiveness of the proposed method was proved by experimental verification in the Jetson Xavier Nx development board and public dataset KITTI.更多还原