Depth completion method applied to single-line laser radar

The invention discloses a depth completion method applied to a single-line laser radar, and the method comprises the steps: processing a KITTI 64 line depth completion data set to obtain a single-line depth completion data set, training a double-branch deep learning network through the single-line depth completion data set, and obtaining a dense depth map through employing an RGB image and an aligned single-line depth map as the input of a neural network. According to the method, a 64-line laser radar is replaced by the single-line laser radar to realize depth completion, the acquisition cost of the dense depth map is remarkably reduced, and the method can be applied to the single-line laser radar and any n-line (n is smaller than or equal to 64) laser radars, such as a 4-line laser radar and a 16-line laser radar. 一种应用于单线激光雷达的深度补全方法，对KITTI 64线深度补全数据集处理得到单线深度补全数据集，并使用单线深度补全数据集训练双分支深度学习网络后，以RGB图与对齐过的单线深度图为神经网络的输入便能得到稠密深度图。本发明通过单线激光雷达来替换64线激光雷达实现深度补全，显著降低稠密深度图的获取成本，不仅能应用于单线激光雷达，同时可以应用于任意n线(n≤64)激光雷达，如4线激光雷达、16线