LiDAR Spatial Map Accuracy via Camera Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating spatial maps using LiDAR scan data alone often result in inaccuracies due to the inability to distinguish between walls and objects in a space, leading to distorted maps.
Innovation Solution
An electronic device that combines LiDAR scan data with captured images to generate a spatial map, performing object detection and depth calibration to accurately reflect the structure of the space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If LiDAR scan data alone is used to generate spatial map, then the generation process is simple and fast, but the accuracy of spatial map is poor due to inability to distinguish walls from objects
Solution Approach 1:
The patent combines LiDAR scan data with camera captured images to generate the spatial map. The LiDAR provides depth information while the camera provides visual information for object recognition. By merging these two data sources, the system can distinguish between walls and objects more accurately than using LiDAR alone, resolving the contradiction between map accuracy and system complexity.
Solution Approach 2:
The patent introduces an intermediary processing step that analyzes captured images to identify objects and uses this information to correct the spatial map generated from LiDAR data. This intermediary image analysis acts as a mediator that enhances the accuracy of wall detection without requiring complete system redesign, thus improving measurement precision while managing device complexity.
2Reliability
If LiDAR scan data alone is used, then the processing time is short, but the spatial map becomes distorted due to misidentification of objects as walls
Solution Approach 1:
The patent performs preliminary object detection and identification using camera images before finalizing the spatial map generation. By提前 identifying objects in the captured images, the system can pre-mark these areas to be excluded from wall detection, thereby improving map reliability. This preliminary action prevents misidentification errors before they affect the final map, balancing reliability improvement with acceptable processing time.
Solution Approach 2:
The patent implements a feedback mechanism where the captured images provide corrective information to the LiDAR-based spatial map. The image data feedbacks on the preliminary map to identify and correct misidentified regions where objects were mistakenly detected as walls. This feedback loop continuously improves map reliability by eliminating distortion caused by object misidentification.
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
The method provides a more accurate spatial map by correctly identifying objects and walls, ensuring that the map accurately represents the actual space, even in the presence of furniture and home appliances.
Implementation Method 1
obtain light detection and ranging (LiDAR) scan data by scanning a depth of the target area with respect to a first height
Data Source
AI summary
A method of generating a spatial map includes obtaining a captured image of a target area in a space, obtaining light detection and ranging (LiDAR) scan data by scanning a depth of the target area with respect to a first height, performing object detection on the captured image, and, according to the performed object detection, generating a spatial map of the target area, based on all of the LiDAR scan data and the captured image.