Autonomous Vehicle Obstacle Detection via LIDAR and Camera Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous vehicles face challenges in detecting obstacles, especially at high speeds and when obstacles are far away, due to inaccuracies in LIDAR data and increased computation time required for camera-based obstacle detection systems.
Innovation Solution
A method and system that combines LIDAR point cloud data and camera images to generate disparity images, extract obstacle regions, determine boundary points, and calculate gradients for accurate obstacle detection, allowing for real-time identification of obstacles without relying on prior frame data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If LIDAR sensors are used for obstacle detection, then detection speed is improved, but measurement precision deteriorates for far distance obstacles
Solution Approach 1:
The patent combines LIDAR sensors and camera sensors into a hybrid obstacle detection system. The LIDAR provides fast detection speed while the camera provides complementary visual information that improves measurement precision for far distance obstacles. The system fuses data from both sensors to achieve both speed and precision requirements.
Solution Approach 2:
The camera sensor serves multiple functions: it provides visual confirmation of obstacles, helps identify obstacle types, and supplements LIDAR data for improved precision. This multi-functionality allows the system to maintain fast detection while enhancing accuracy through the camera's additional capabilities.
2Measurement precision
If camera imaging is used for obstacle detection, then measurement precision is improved, but computation time increases
Solution Approach 1:
By merging camera and LIDAR systems, the patent leverages the camera's high precision obstacle detection while using LIDAR's faster processing capabilities to compensate for the camera's longer computation time. The LIDAR provides rapid initial detection that reduces the computational burden on the camera processing system.
Solution Approach 2:
The system processes only the most critical image data from the camera rather than all available data, using LIDAR to pre-identify potential obstacle regions. This partial processing approach maintains measurement precision while significantly reducing computation time by focusing resources only on areas where obstacles are detected by LIDAR.
3Measurement precision
If multiple frames of images are processed to detect obstacles, then measurement precision is improved, but productivity decreases
Solution Approach 1:
The hybrid system combines LIDAR's real-time detection capability with camera's precision. LIDAR processes data continuously without requiring multiple frames, while the camera provides supplemental precision. This merging allows the system to achieve both high productivity and measurement precision simultaneously.
Solution Approach 2:
LIDAR performs preliminary obstacle detection before camera processing begins. By pre-identifying potential obstacles through LIDAR, the system can then process only the most relevant camera frames, reducing the total number of frames that need processing while maintaining high measurement precision. This preliminary action significantly improves productivity.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables accurate and efficient detection of obstacles in real-time, even at high speeds and long distances, providing precise location, structure, and orientation information for obstacle maneuvering.
Implementation Method 1
receiving, by an Electronic Control Unit (ECU) of a vehicle, one or more images of surroundings of the vehicle from an imaging unit and point cloud data of the surroundings of the vehicle from a Light Detection and Ranging (LIDAR) unit
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure relates to detection of obstacles by an autonomous vehicle in real-time. An obstacle detection system of an autonomous vehicle obtains point cloud data from single frame Light Detection and Ranging (LIDAR) data and camera data, of the surroundings of the vehicle. Further, the system processes the camera data to identify and extract regions comprising the obstacles. Further, the system extracts point cloud data corresponding to the obstacles and enhances the point cloud data, with the detailed information of the obstacles provided by the camera data. Further, the system processes the enhanced point cloud data to determine the obstacles along with the structure and orientation of obstacles. Upon determining the details of the obstacles, the system provides instructions to the vehicle to manoeuvre the obstacle. The disclosed obstacle detection system provides accurate data about a structure, an orientation and a location of the obstacles.