Autonomous Driving Solutions

Autonomous Driving Solutions

VendeTech provides systematic, diversified and multi-leveled data services for multiple application scenarios of intelligent driving, from vehicle outside environmental perception to intelligent interaction in the cockpit, as well as the basic roadside requirements of high-precision maps.

Solution Architecture

Through our powerful data processing platform, AI preprocessing service and combined data labeling service service, a data service solution for intelligent driving is formed, providing efficient and high-quality services to help customers train AI models.In terms of environmental perception, we provide 2D and 3D full-scene annotations, including 2D images from cameras, lidar point clouds, millimeter-wave radar obstacles, and other multi-channel sensor fusion annotations. As for smart cockpit, we provide rich human-machine voice interaction and driving behavior tracking & annotation. We also provide high-precision maps for the road end, including static street view maps, dynamic road information, and 3D street view modeling services.

Panoramic Semantic Segmentation

Use polygons, points, lines, etc. to accurately mark all objects in the image, including roads and facilities, obstacles, trees, buildings, etc.

Obstacles Labeling

Various obstacles on the road, including vehicles, bicycles, pedestrians and other driving obstacles, are framed and marked. Image and video continuous frame intelligent auxiliary annotation is supported

Drivable Area Labeling

Semantic labeling of lane lines, zebra crossings, parking spaces, and drivable areas on the road

Image class

3D Point Cloud

3D Point Cloud Object Detection

in the laser radar 3D point cloud, accurate annotation of various types of motor vehicles, riding, pedestrians and other movable objects.

2D/3D Combined Annotation

According to the radar and camera multi-sensor data, complete the combined annotation of 3D point cloud and 2D images, and automatically complete the mapping relationship between 3D and 2D

3D Point Cloud Semantic Segmentation

Data annotation such as panoramic semantic segmentation and entity segmentation is completed in the lidar 3D point cloud.

Environmental Awareness

For the application of laser radar and visual fusion perception in autonomous driving, most high-level intelligent driving passenger cars choose to use different sensors for fusion, and their advantages are complementary. VendeTech participates in the fusion of point clouds and images in the direction of environmental perception and multi-sensor fusion annotation.

Vehicular Speech

Collect in-vehicle voice conversations and audio, process and label conversations, sounds, and noises in specific scenarios; help customers quickly build interactive smart cockpits with rich scenes. Multilingual and multi-environmental wake-up words and command language collection are supported.

DMS

Collect and mark images such as driver's facial expression and driver's behavior, reduce the accident rate of vehicle driving by monitoring driver's fatigue state and dangerous actions in the cockpit, etc., providing support for DMS research and development.

Smart Cockpit

VendeTech serves various scenarios in the smart cockpit, of which the design requirements are diverse: In addition to having the ability to process massive amounts of information and data, it is also necessary to improve the efficient and intuitive driving experience for users and ensure that users drive safe and comfortable

HD Map

HD Map contains high-density road information including road geometry, slope, curvature, heading, detailed lane information and road surrounding pair information, and constructs an accurate driving environment for the vehicle. Advanced sensing equipment is combined with it to realize the automatic driving of the vehicle.

Static Data

Update monthly Including: lane-level road topology data, road infrastructure data, and road infrastructure attribute data.

Quasi-static data

Including: traffic sign location and meaning, traffic light location, point of interest (POI) location, and landmark building location. Among them, traffic signs and traffic lights belong to the roadside display equipment. Traffic signs include: special lanes, lane directions, vehicle speed limits, traffic restrictions, curve warnings, etc.

Dynamic Data

The first type is traffic operation data such as traffic condition, road performance and road weather; the second type is traffic management data such as temporary traffic sign and traffic control.