Contributions:
First, we demonstrate the crucial effects of data diversity on the generalization ability of a VO model by comparing performance on different quantities of training data. (deeplearning 的痛点 误差找到姿态变化丰富的数据集,作者采用合成的办法,可以提供各种各样的变换姿态的GT)
Second, we design an up-to-scale loss function to deal with the scale ambiguity of monocular VO. (通过归一化的方式,解决尺度歧义的问题)Third, we create an intrinsics layer (IL) in our VO model enabling generalization across different cameras. (提出了一个内参层,解决训练集合测试集内参不一样造成的问题)