自动驾驶合成数据数据集

<p>LightwheelOcc 自动驾驶合成数据数据集</p><p>数据来源：公开网站</p><p>数据规模：40,000帧及其对应的各种任务的真值标签</p><p>数据特点：合成数据概括性数据集，覆盖了多种地区地形、天气模式、车辆类型、植被和道路标记。LightwheelOcc 数据集包含 40,000 帧，共计 240,000 张图像，其中 28,000 帧用于训练场景，6,000 帧用于验证场景，6,000 帧用于测试场景。LightwheelOcc 包括 6 个相机传感器的数据，以及用于不同任务的标签，包括 3D 占用、流和深度图。</p><p>应用场景：自动驾驶汽车训练</p><p>LightwheelOcc: Synthetic Autonomous Driving Dataset</p><p>Data Source: Publicly Available Websites</p><p>Dataset Scale: 40,000 frames with corresponding ground truth labels for various tasks</p><p>Dataset Characteristics: This is a synthetic general-purpose dataset covering diverse regional terrains, weather patterns, vehicle types, vegetation and road markings. The LightwheelOcc dataset contains 40,000 frames, totaling 240,000 images, with 28,000 frames allocated for training scenarios, 6,000 for validation scenarios, and 6,000 for testing scenarios. It includes data from 6 camera sensors, as well as labels for different tasks, including 3D occupancy, optical flow and depth maps.</p><p>Application Scenario: Autonomous Vehicle Training</p>