上海科技大学知识管理系统

本发明的一个技术方案是提供了一种高能效FPGA实现的点云特征提取方法，映射到FPGA上运行。本发明的另一个技术方案是提供了一种前述的高能效FPGA实现的点云特征提取方法的应用，其特征在于，应用于无人驾驶或者机器人的点云特征提取中。相比于现有技术方案而言，本发明的创新之处在于：一种低复杂度的投影方法来组织无序且稀疏的点云；一种高并行度的方法来提取粗粒度特征点；一种高并行度的方法来选择细粒度的特征点。

One technical solution of the present invention provides a high-energy-efficiency FPGA-based point cloud feature extraction method, which is mapped to an FPGA for running. Another technical solution of the present invention provides a use of the high-energy-efficiency FPGA-based point cloud feature extraction method, wherein the method is used for point cloud feature extraction of unmanned driving or robots. Compared with the prior art, the innovation of the present invention is : providing a low-complexity projection method for organizing disordered and sparse point clouds, a high parallelism method for extracting coarse-grained feature points, and a high parallelism method for selecting fine-grained feature points.