[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-93435":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":9,"totalLinesOfCode":9,"stars":12,"forks":13,"watchers":14,"openIssues":14,"contributorsCount":9,"subscribersCount":14,"size":14,"stars1d":14,"stars7d":14,"stars30d":14,"stars90d":14,"forks30d":14,"starsTrendScore":14,"compositeScore":15,"rankGlobal":9,"rankLanguage":9,"license":9,"archived":16,"fork":16,"defaultBranch":17,"hasWiki":16,"hasPages":16,"topics":9,"createdAt":9,"pushedAt":9,"updatedAt":18,"readmeContent":19,"aiSummary":20,"trendingCount":14,"starSnapshotCount":14,"syncStatus":21,"lastSyncTime":9,"discoverSource":22},93435,"vision_pilot","autowarefoundation\u002Fvision_pilot","autowarefoundation","Free and fully open-source L2 ADAS stack powered by End-to-End AI technology",null,"https:\u002F\u002Fgithub.com\u002Fautowarefoundation\u002Fvision_pilot","C++",669,146,0,47.5,false,"main","2026-07-22 04:02:09","# Vision Pilot - Open Source L2 ADAS\n\n\u003Cp align=\"center\">\n    \u003Cpicture>\n        \u003Csource media=\"(prefers-color-scheme: dark)\">\n        \u003Cimg src=\".\u002FMedia\u002FVisionPilot_logo.png\" alt=\"VisionPilot\" width=\"100%\">\n    \u003C\u002Fpicture>\n\u003C\u002Fp>\n\n\u003Cdiv align=\"center\">\n\n[![License](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-Apache_2.0-blue.svg)](https:\u002F\u002Fopensource.org\u002Flicenses\u002FApache-2.0)\n[![Discord](https:\u002F\u002Fimg.shields.io\u002Fdiscord\u002F953808765935816715?label=Autoware%20Discord)](https:\u002F\u002Fdiscord.com\u002Finvite\u002FQ94UsPvReQ)\n![GitHub commit activity](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fcommit-activity\u002Fm\u002Fautowarefoundation\u002Fautoware.privately-owned-vehicles)\n![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fautowarefoundation\u002Fautoware.privately-owned-vehicles)\n\n![PyTorch](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPyTorch-EE4C2C?style=for-the-badge&logo=pytorch&logoColor=white)\n![OpenCV](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FOpenCV-27338e?style=for-the-badge&logo=OpenCV&logoColor=whit)\n![ROS](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FROS-22314E?style=for-the-badge&logo=ROS&logoColor=whit)\n[![LinkedIn](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLinkedIn-0077B5?style=for-the-badge&logo=linkedin&logoColor=white)](https:\u002F\u002Fwww.linkedin.com\u002Fcompany\u002Fthe-autoware-foundation)\n[![YouTube](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FYouTube-FF0000?style=for-the-badge&logo=youtube&logoColor=white)](https:\u002F\u002Fwww.youtube.com\u002F@autowarefoundation)\n[![Website](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fwebsite-000000?style=for-the-badge&logo=About.me&logoColor=white)](https:\u002F\u002Fautoware.org\u002F)\n\u003C\u002Fdiv>\n\n\u003Cdiv align=\"center\">\n\n⭐ Star us on GitHub — your support motivates us a lot!\n\n\u003C\u002Fdiv>\n\n## Free and fully open-source stack for L2 ADAS\n[![Watch the Video](\u002FMedia\u002FVision_Pilot_Thumbnail.jpg)](https:\u002F\u002Fdrive.google.com\u002Ffile\u002Fd\u002F1pAxpppljBdDKFYgrKdWiwUtPiEBYbVhL\u002Fview)\n\n**This codebase contains a productionizable and safety certifiable implementation of an open-source L2 ADAS system called Vision Pilot.**\n\nVision Pilot is designed to be integrated with automotive OEMs and Tier-1 suppliers in series production passenger cars, and the system can optionally be adopted for transportation and logisitcs use-cases in buses and trucks. \n\nWe offer the complete codebase as free and fully open-source, including AI model weights to help democratize access to this vital technology. Vision Pilot is available under the persmissive Apache 2.0 licence and can freely be used for both commercial and research purposes.\n\n\n\u003Cimg src=\"\u002FMedia\u002FVisionPilot_1.0.png\" width=\"100%\">\n\nVision Pilot is designed to support basic\u002Fentry L2 ADAS features for in-lane autonomous driving including the following features:\n\n- **ACC** - autonomous cruise control\n- **FCW** - forward collision warning\n- **AEB** - autonomous emergency braking\n- **LKAS** - lane keep assist\n- **LDW** - lane departure warning\n- **ISA** - intelligent speed assist\n- **Autopilot** - single-lane hands-free highway autopilot\n\n**Sensor specification:**\n\nVision Pilot can be run with a single, front-facing, monocular camera with 52 - 55 degree horizontal field-of-view, and 1MP - 2MP resolution.\n\n### Hybrid End-to-End AI Architecture\nWe utilize a **Hybrid End-to-End AI Architecture** as the core of Vision Pilot in which data is processed in parallel by perception AI models for safety, and End-to-End AI models for performance.\n\n\u003Cimg src=\"\u002FMedia\u002FVisionPilot_architecture.png\" width=\"100%\">\n\nSpecifically, Vision Pilot leverages three open-source AI models developed by the Autoware Foundation, these are:\n\n1. [AutoSpeed](https:\u002F\u002Fgithub.com\u002Fautowarefoundation\u002Fauto_speed) - Closest in-path object detection\n2. [AutoSteer](https:\u002F\u002Fgithub.com\u002Fautowarefoundation\u002Fauto_steer) - Ego path future waypoint detection\n3. [AutoDrive](https:\u002F\u002Fgithub.com\u002Fautowarefoundation\u002Fauto_drive) - End to End distance\u002Fin-path object presence detection and road curvature estimation\n\n### No Reliance on High Definition Maps\n**Vision Pilot does not require 3D high definition maps** and operates in a 'mapless' mode to follow the road in real-time.\n\n### Presentation\nFor more information about Vision Pilot, please view our introductory presentation at this [link](https:\u002F\u002Fcanva.link\u002Fqjz6rwp40v7apft)\n\n## Getting Started\n\nThere are three  ways you can build and run Vision Pilot - we provide example data which you can use for visualization and evaluation, select from the drop down options below:\n\n\u003Cdetails>\n\u003Csummary>\u003Cb>Option 1 - Download and build from source\u003C\u002Fb>\u003C\u002Fsummary>\n\nTo get started with the project, download the source code from:\n```bash\n  git clone https:\u002F\u002Fgithub.com\u002Fautowarefoundation\u002Fautoware_vision_pilot.git\n```\nDownload ONNX Runtime from the GitHub [releases](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fonnxruntime\u002Freleases) page.\n\nBuild the project:\n\n```bash\n  cd VisionPilot\n```\n\n```bash\n  mkdir build && cd build\n```\n\n```bash\n  cmake -DONNXRUNTIME_ROOT=\u003CONNX_RUNTIME_ROOT_PATH> ..\u002F\n```\n\nor with ROS2 support \n\n```bash\n  cmake -DONNXRUNTIME_ROOT=\u003CONNX_RUNTIME_ROOT_PATH> -DENABLE_ROS2_INTERFACE=ON ..\u002F\n```\n\n```bash\n  make\n```\nThis will build the project and create VisionPilot executable inside the build directory.\n\n#### Run Vision Pilot on test data and visualize outputs\n\n**OpenLane Dataset:**\n\nTo test Vision Pilot using open loop scenario testing, first download the sample data from the [Google Drive](https:\u002F\u002Fdrive.google.com\u002Fdrive\u002Ffolders\u002F1-Sxgz3XHzFD6XtETz1sVFRtDKY3W57QB?usp=sharing) directory.\n\nThis directory contains video composed of image data from sequences in the appropriate dataset and vehicle speed data \nextracted from the dataset.\n\nUpdate VisionPilot config files `vision_pilot.conf` inside `config` directory and set: \n\n```\nsource.mode             = video\n```\nand `vision_pilot_test.conf` set:\n\n```\nsource.input_video         = \u003CINPUT_VIDEO_FILE_PATH>\nsource.input_vehicle_speed = \u003CINPUT_VEHICLE_SPEED_FILE_PATH>\n```\nto point to the appropriate video file path and vehicle speed file path.\n\n*Note*: When VisionPilot built from source, update config files before the build. \n\nRun VisionPilot from inside `build` directory\n\n```\n.\u002FVisionPilot\n```\n\n#### Build VisionPilot DEB package\n\nTo build Debian DEB package for VisionPilot run \n\n```bash\n  cpack -G DEB\n```\nTo build Debian DEB package for VisionPilot with CPU support build VisionPilot with \n\n```bash\n  cmake -DONNXRUNTIME_ROOT=\u003CONNX_RUNTIME_ROOT_PATH> -DGPU=OFF ..\u002F\n```\n```bash\n  make\n```\n```bash\n  cpack -G DEB\n```\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cb>Option 2 - Use the pre-built Debian pacakge\u003C\u002Fb>\u003C\u002Fsummary>\n\n### Install from the pre-built Debian package\n\nThis method is recommended in case new system installation and cuda dependencies are not installed yet.\n\nDownload [VisionPilot](https:\u002F\u002Fgithub.com\u002Fautowarefoundation\u002Fautoware_vision_pilot.git) prebuilt binary.\n\nInstall the .deb package\n\n```bash\n  sudo apt install .\u002FVisionPilot-1.0-x86_64.deb\n```\nReboot the system, cuda dependencies for VisionPilot installed.\n\n#### Run Vision Pilot on test data and visualize outputs\n\n**OpenLane Dataset:**\n\nTo test Vision Pilot using open loop scenario testing, first download the sample data from the [Google Drive](https:\u002F\u002Fdrive.google.com\u002Fdrive\u002Ffolders\u002F1-Sxgz3XHzFD6XtETz1sVFRtDKY3W57QB?usp=sharing) directory.\n\nThis directory contains video composed of image data from sequences in the appropriate dataset and vehicle speed data \nextracted from the dataset.\n\nUpdate VisionPilot config files `vision_pilot.conf` set: \n\n```\nsource.mode             = video\n```\nand `vision_pilot_test.conf` set:\n\n```\nsource.input_video         = \u003CINPUT_VIDEO_FILE_PATH>\nsource.input_vehicle_speed = \u003CINPUT_VEHICLE_SPEED_FILE_PATH>\n```\nto point to the appropriate video file path and vehicle speed file path.\n\n*Note*: When VisionPilot installed from prebuilt binary update the config files inside: \n\n```\n\u002Fusr\u002Fshare\u002Fvisionpilot\u002Fconfig\n```\ndirectory.\n\nRun VisionPilot from the command line\n\n```\nVisionPilot\n```\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cb>Option 3 - Use the Docker container\u003C\u002Fb>\u003C\u002Fsummary>\n\n### Run inside Docker container\n\nTo run Vision Pilot inside a Docker container, build the container using the Dockerfiles provided in the docker directory of the repository.\nDocker containers can be built with GPU\u002FCPU support, and NO_ROS2\u002FROS2 support.\n\nTo build the container, go to the docker subdirectory and run the following commands:\n\nDefault is GPU support NO ROS2 support\n\n```bash\n  .\u002Fbuild.sh --gpu --ros2 \n```\nto build with CPU support\n\n```bash\n  .\u002Fbuild.sh --cpu\n```\n\nTo run the container use the `run.sh` script. For example to run the container with CPU support \n\n```bash\n  .\u002Frun.sh --cpu\n```\n\nor with GPU support and ROS2 support\n\n```bash\n  .\u002Frun.sh --gpu --ros2\n```\n\n*_Note:_* Before building the container update the config files inside: `config` directory.\n\nIf building with CPU support update `config\u002Fvision_pilot.conf`\n```\nengine.provider     = cpu\n```\nand when building with ros2 support update \n```\nsource.mode         = ros2\n```\nWhen using video input source update `config\u002Fvision_pilot_test.conf`\n```\nsource.input_video         = \u003CINPUT_VIDEO_FILE_PATH>\nsource.input_vehicle_speed = \u003CINPUT_VEHICLE_SPEED_FILE_PATH>\n```\nInput directory can be modified in the source by using `run.sh` script switch\n\n```bash\n  .\u002Frun.sh --gpu --data \u003CHOST_DIR>:\u003CCONTAINER_DIR>\n```\n*_Note:_* \u003CCONTAINER_DIR> need to be the same as during the container build.\n\n\u003C\u002Fdetails>\n\n## Running Vision Pilot with your own camera\nIn order to run Vision Pilot with your own camera, you will need to calibrate your camera and provide the calibraiton information to the Vision Pilot app - this is important for Vision Pilot to accurately measure the shape of the road and the distance to objects.\n\n### Calibrate your camera\nPlease follow the steps in the [calibration guide](\u002FCalibration\u002F) and save the calibration yaml file in the [Vision Pilot config folder](\u002FVisionPilot\u002Fconfig\u002F) and overwrite the data of the [H.yaml](\u002FVisionPilot\u002Fconfig\u002FH.yaml) file with your calibraiton file parameters - we suggest you retain a copy of the original H.yaml parameters so that you can run Vision Pilot on our example data.\n\n## Roadmap\n- Guide on how to run Vision Pilot with simulators\n- Support for 8MP camera resolution with 120 degree horizontal field-of-view\n- Support for fusion between front-facing camera and automotive RADAR\n- Safety Verification and Automotive Standards Compliance (ISO26262, ISO8800)\n\n## Contributing\n\nTo learn more about how to participate in this project, please read the [onboarding guide](\u002FONBOARDING.md)","Vision Pilot 是一个免费且完全开源的 L2 级高级驾驶辅助系统（ADAS）软件栈，基于端到端 AI 技术构建。核心功能包括自适应巡航（ACC）、前向碰撞预警（FCW）、自动紧急制动（AEB）、车道保持辅助（LKAS）、车道偏离预警（LDW）、智能限速辅助（ISA）及单车道高速领航（Autopilot），仅需单目前视摄像头即可运行。技术上融合 PyTorch 模型推理、OpenCV 图像处理与 ROS 中间件，支持车规级集成与功能安全认证。适用于乘用车量产开发，亦可拓展至客车与商用车的辅助驾驶场景。",2,"trending"]