awesome-autopilot-algorithm
github.com/hardy-uint/awesome-autopilot-algorithm ↗some algorithm about self-driving car,mainly including perception algorithm,2D/3D object detection,Semantic segmentation and so on
Use this list with your AI agent
Add the Context Awesome MCP server to Claude, Cursor, or any MCP client, then ask:
"Show me 目标检测 resources from awesome-autopilot-algorithm"
Installation instructions →What's inside
感知
- 3D Object Detection Using Scale Invariant and Feature Reweighting Networks目标检测
- 3DSSD: Point-based 3D Single Stage Object Detector(CVPR2020)目标检测
- Accurate Monocular Object Detection via Color-Embedded 3D Reconstruction for Autonomous Driving目标检测
- BirdNet: a 3D Object Detection Framework from LiDAR information目标检测
- Class-balanced Grouping and Sampling for Point Cloud 3D Object Detection目标检测
- Class-specific Anchoring Proposal for 3D Object Recognition in LIDAR and RGB Images目标检测
传感器标定融合
- A Novel Method for the Extrinsic Calibration of a 2D Laser Rangefinder and a Camera数据融合
- A Perceptual Measure for Deep Single Image Camera Calibration相机在线标定
- Automatic extrinsic calibration between a camera and a 3D Lidar using 3D point and plane correspondences数据融合
- CalibNet: Geometrically Supervised Extrinsic Calibration using 3D Spatial Transformer Networks传感器标定
- Extrinsic camera calibration method and its performance evaluation相机在线标定
- Forward Vehicle Collision Warning Based on Quick Camera Calibration相机在线标定
Resources
- Apollo
- Autoware
- Caltech数据集
加州理工学院行人数据集包括大约10小时的640x480 30Hz视频,这些视频来自在城市环境中通过常规交通的车辆。 大约250,000个帧(137个近似分钟的长段)共有350,000个边界框和2300个独特的行人被注释。 注释包括边界框和详细遮挡标签之间的时间对应。 更多信息可以在我们的PAMI 2012和CVPR 2009基准测试文件中找到。
- CamVid
- Cityscape Dataset
专注于对城市街景的语义理解。 大型数据集,包含从50个不同城市的街景中记录的各种立体视频序列,高质量的像素级注释为5000帧,另外还有一组较大的20000个弱注释帧。 因此,数据集比先前的类似尝试大一个数量级。 可以使用带注释的类的详细信息和注释示例。
- Comma.ai
7.25小时的高速公路驾驶。 包含10个可变大小的视频片段,以20 Hz的频率录制,相机安装在Acura ILX 2016的挡风玻璃上。与视频平行,还记录了一些测量值,如汽车的速度、加速度、转向角、GPS坐标,陀螺仪角度。 这些测量结果转换为均匀的100 Hz时基。
Showing a sample of 180 resources. View the full list on GitHub →