Real-time Range Map Generation for Single Camera Visual Odometry
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing visual odometry systems for heavy equipment, such as those described in U.S. Pat. No. 8,238,612, are inadequate for machines with a single camera and non-uniform terrain, as they require stereo images and cannot effectively fuse LIDAR data with single-camera image data to generate accurate range maps.
Innovation Solution
A range map generation system that includes a camera, a LIDAR unit, and a controller, which receives images and 3D point clouds, transforms them into camera coordinates, projects them into a 2D image space, filters points based on a range threshold, and generates a range map combining camera and LIDAR data to estimate motion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a single camera is used instead of stereo cameras or LIDAR, then cost is reduced, but the ability to generate accurate range maps and determine motion is degraded
Solution Approach 1:
The patent combines data from a single camera with LIDAR range data to create a fused range map. The camera provides image data while the LIDAR provides depth information, and these are merged through coordinate transformation and projection to produce an enhanced range map that overcomes the limitations of using either sensor alone.
Solution Approach 2:
The patent uses an intermediary coordinate transformation process to bridge the camera image space and LIDAR range space. By transforming LIDAR points into camera coordinates and projecting them onto the image plane, the system creates a unified representation that enables accurate range map generation from a single camera.
2Measurement precision
If LIDAR data is fused with single camera image data, then range map accuracy is improved, but system complexity increases
Solution Approach 1:
The patent segments the fusion process into distinct computational stages: coordinate transformation of LIDAR points, projection onto the camera image plane, filtering based on range thresholds, and final range map generation. This segmentation makes the complex fusion process more manageable and implementable.
3Manufacturing precision
If real-time range map generation is implemented, then operational precision is improved, but processing requirements and system complexity increase
Solution Approach 1:
The patent performs preliminary coordinate transformation and projection of LIDAR points before the actual range map generation. By pre-processing the LIDAR data into camera coordinates and projecting it onto the image plane in advance, the system reduces the computational burden during real-time operation.
Solution Approach 2:
The patent applies range threshold filtering to change the parameters of the projected points, keeping only those within a valid range. This parameter-based filtering simplifies the data structure and reduces computational complexity for subsequent processing steps.
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 real-time range map generation and motion estimation for heavy equipment, even with a single camera, improving precision and safety in autonomous operations on varying terrains by integrating camera and LIDAR data effectively.
Implementation Method 1
A LIDAR (light detection and ranging) unit 36 may be used to obtain a 3D point cloud image
Data Source
AI summary
A system, method, and non-transitory computer-readable storage medium for range map generation is disclosed. The method may include receiving an image from a camera and receiving a 3D point cloud from a range detection unit. The method may further include transforming the 3D point cloud from range detection unit coordinates to camera coordinates. The method may further include projecting the transformed 3D point cloud into a 2D camera image space corresponding to the camera resolution to yield projected 2D points. The method may further include filtering the projected 2D points based on a range threshold. The method may further include generating a range map based on the filtered 2D points and the image.