摘要
由一名新闻记者-机器人与机器学习每日新闻编辑-研究人员详细介绍了机器人S-机器人和自动化的新数据。根据NewsRx记者从德国乌尔姆发回的消息，研究表明：“不利的天气条件会通过在测量中引入不想要的噪声而严重影响激光雷达传感器的性能。因此，区分噪声和有效点是这些传感器可靠使用的关键。”我们的新闻记者从乌尔姆大学的研究中获得了一句话：“目前探测不利天气点的方法需要大量的标签数据，”本文提出了一种基于标签的LiDAR点云数据分割方法,提出了一种利用少量镜头语义分割的LiDAR点云数据分割框架,并利用半监督学习方法对未标记点云数据进行伪标签生成。在不需要任何传统标签的情况下，显著增加了训练数据量。我们还将良好的天气数据整合到训练管道中，允许在良好和不利的天气条件下都有很高的性能。真实和合成数据集的实验结果表明，我们的方法在探测雪、雾和喷雾方面表现良好。
Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News – Researchers detail new data in Robotic s - Robotics and Automation. According to news originating from Ulm, Germany, by NewsRx correspondents, research stated, “Adverse weather conditions can severel y affect the performance of LiDAR sensors by introducing unwanted noise in the m easurements. Therefore, differentiating between noise and valid points is crucia l for the reliable use of these sensors.” Our news journalists obtained a quote from the research from Ulm University, “Cu rrent approaches for detecting adverse weather points require large amounts of l abeled data, which can be difficult and expensive to obtain. This letter propose s a label-efficient approach to segment LiDAR point clouds in adverse weather. W e develop a framework that uses few-shot semantic segmentation to learn to segme nt adverse weather points from only a few labeled examples. Then, we use a semi- supervised learning approach to generate pseudo-labels for unlabelled point clou ds, significantly increasing the amount of training data without requiring any a dditional labeling. We also integrate good weather data in our training pipeline , allowing for high performance in both good and adverse weather conditions. Res ults on real and synthetic datasets show that our method performs well in detect ing snow, fog, and spray.”
NETL
NSTL