Centre for Computational Engineering Sciences (CES), School of Aerospace, Transport and Manufacturing (SATM), Cranfield University, Bedfordshire MK43 0AL, UK.
Centre for Life-Cycle Engineering and Management, School of Aerospace, Transport and Manufacturing (SATM), Cranfield University, Bedfordshire MK43 0AL, UK.
Sensors (Basel). 2021 Oct 21;21(21):6996. doi: 10.3390/s21216996.
Sky and ground are two essential semantic components in computer vision, robotics, and remote sensing. The sky and ground segmentation has become increasingly popular. This research proposes a sky and ground segmentation framework for the rover navigation visions by adopting weak supervision and transfer learning technologies. A new sky and ground segmentation neural network (network in U-shaped network (NI-U-Net)) and a conservative annotation method have been proposed. The pre-trained process achieves the best results on a popular open benchmark (the Skyfinder dataset) by evaluating seven metrics compared to the state-of-the-art. These seven metrics achieve 99.232%, 99.211%, 99.221%, 99.104%, 0.0077, 0.0427, and 98.223% on accuracy, precision, recall, dice score (F1), misclassification rate (MCR), root mean squared error (RMSE), and intersection over union (IoU), respectively. The conservative annotation method achieves superior performance with limited manual intervention. The NI-U-Net can operate with 40 frames per second (FPS) to maintain the real-time property. The proposed framework successfully fills the gap between the laboratory results (with rich idea data) and the practical application (in the wild). The achievement can provide essential semantic information (sky and ground) for the rover navigation vision.
天空和地面是计算机视觉、机器人技术和遥感中的两个基本语义组件。天空和地面分割已经变得越来越流行。本研究通过采用弱监督和迁移学习技术，为漫游者导航视觉提出了一个天空和地面分割框架。提出了一种新的天空和地面分割神经网络（U 型网络中的网络（NI-U-Net））和一种保守的注释方法。通过评估七个指标，与最先进的方法相比，在流行的开放基准（Skyfinder 数据集）上，预训练过程实现了最佳结果。这七个指标在准确性、精度、召回率、骰子分数（F1）、误分类率（MCR）、均方根误差（RMSE）和交并比（IoU）上分别达到 99.232%、99.211%、99.221%、99.104%、0.0077%、0.0427%和 98.223%。保守的注释方法在有限的人工干预下可以实现优异的性能。NI-U-Net 可以以每秒 40 帧（FPS）的速度运行，以保持实时性。所提出的框架成功填补了实验室结果（具有丰富的想法数据）和实际应用（在野外）之间的差距。该成果可为漫游者导航视觉提供必要的语义信息（天空和地面）。
