Mobile Mapping Object Detection via Laser-Image Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for collecting and processing terrestrial mobile mapping data, particularly high-resolution laser scanner data, are inefficient due to high computational costs and limited suitability for detecting objects by form, shape, or position, and require extensive processing power to analyze vast amounts of data.
Innovation Solution
A method that filters laser scanner data based on vehicle position and orientation to identify regions of interest, maps these points to image coordinates, generates recognition masks, and combines them with source images to create candidate 3D images, reducing data processing time and improving object detection accuracy by incorporating depth information and perpendicularity analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high resolution laser scanner data is used to detect objects, then detection precision is improved, but device cost and processing complexity increase significantly
Solution Approach 1:
The patent segments the laser scanner data into regions of interest based on spatial filtering, separating relevant object data from irrelevant background data. This segmentation allows processing only the necessary portions of data, reducing overall processing complexity while maintaining detection precision for objects within the road corridor.
Solution Approach 2:
The patent extracts and filters only the essential laser points that correspond to regions of interest (road signs, obstacles, etc.) from the vast amount of laser scanner data. By taking out only the relevant data points and projecting them to image coordinates, the system reduces processing complexity while preserving detection accuracy.
2Reliability
If all laser scanner data is processed to detect objects, then detection completeness is improved, but processing time increases significantly
Solution Approach 1:
The patent performs preliminary filtering of laser scanner data before object detection, pre-identifying regions of interest based on spatial criteria. This preliminary action reduces the volume of data that requires full processing, thereby reducing processing time while maintaining detection completeness for objects within the road corridor.
Solution Approach 2:
The patent applies different processing quality levels to different regions of the data. High processing quality is applied only to laser points within the road corridor (regions of interest), while lower processing quality or no processing is applied to points outside the corridor. This local quality approach reduces overall processing time while maintaining detection completeness for relevant objects.
3Loss of time
If laser scanner data is filtered to reduce processing load, then processing time is reduced, but detection accuracy may deteriorate
Solution Approach 1:
The patent uses image coordinates and recognition masks as intermediary structures to bridge the filtered laser data and object detection. By mapping filtered laser points to image coordinates and creating recognition masks, the system maintains detection accuracy even with reduced data volume, as the intermediaries preserve spatial relationships and object characteristics.
Solution Approach 2:
The patent transforms filtered laser points from 3D space to 2D image coordinates through projection. This dimensional change allows the system to process reduced laser data while maintaining detection accuracy, as the projected 2D image data preserves the essential visual characteristics needed for object recognition.
Data Source
AI summary
A method of detecting objects from terrestrial based mobile mapping data is disclosed, wherein the terrestrial based mobile mapping data has been captured by way of a terrestrial based mobile mapping vehicle driving on a road having a driving direction, the mobile mapping data including laser scanner data, source images obtained by at least one camera and position and orientation data of the vehicle, wherein the laser scanner data includes laser points, each laser point having associated position and orientation data, and each source image comprises associated position and orientation data. In at least one embodiment, the method includes: retrieving a position and orientation of the vehicle; filtering the laser scanner data in dependence of the position and orientation of the vehicle to obtain laser points corresponding to regions of interest; retrieving a source image associated with the position and orientation of the vehicle; mapping the laser points corresponding to regions of interest to image coordinates of the source image to generate a recognition mask; combining the recognition mask and the source image to obtain candidate 3D images representative of possible objects within the regions of interest; and, detecting a group of objects from the candidate 3D images. By combining image recognition and laser scanner recognition the detection rate can be increased to a very high percentage, thereby substantially reducing human effort. Furthermore, the generating of regions of interest in the laser data, enables a significant reduction of the processing power and/or the processing time needed to detect the objects in the images.