TinyTaxiNet

该数据集名为TinyTaxiNet系列，它包含了用于基于视觉的自主滑行感知网络，这些网络能够预测飞机在滑行道上的位置和方向。输入数据是对滑行道灰度图像的下采样处理，输出包括横向距离和航向角误差。在此数据集上，我们评估了两套网络，任务是评估感知模型在局部对抗性环境下的鲁棒性。The dataset series is named TinyTaxiNet. It is tailored for visual-based autonomous taxiing perception networks that can predict the position and orientation of aircraft on taxiways. The input data comprises downsampled grayscale images of taxiways, and the outputs include lateral distance and heading angle error. On this dataset, we evaluated two sets of networks, with the task of assessing the robustness of perception models under local adversarial environments.