Mining Vehicle Point Cloud Filtering for Self-Detection Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mobile mining vehicles equipped with sensors like lidar face issues where their structural body members are inadvertently detected as obstacles, leading to false alarms and missed obstacle detection due to self-interference in sensor data, affecting mapping and obstacle detection algorithms.
Innovation Solution
An environment-related data management apparatus for mobile mining vehicles that obtains real-time position and space information of structural body members, associates bounding boxes with them, and generates an overlap indication to exclude overlapping data points from point cloud data received from sensors, ensuring accurate obstacle detection and mapping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sensors are installed on mobile mining vehicles to detect environment, then obstacle detection capability is improved, but the vehicle's own structural body members are detected as obstacles causing false alarms
Solution Approach 1:
The patent extracts and removes the harmful self-detection data from the sensor point cloud by identifying points that fall within the vehicle's bounding box volume. This separation isolates the useful environmental obstacle detection data from the harmful self-interference data, resolving the false alarm problem while maintaining detection accuracy.
2Speed
If sensors scan the environment continuously, then real-time obstacle detection is improved, but computational complexity increases due to processing all point cloud data
Solution Approach 1:
The patent segments the point cloud data processing into two distinct stages: first filtering points within the vehicle's bounding box (self-interference removal), then processing the remaining points for obstacle detection. This segmentation reduces the volume of data requiring complex obstacle detection algorithms, thereby lowering computational complexity while maintaining real-time processing capability.
3Measurement precision
If mapping algorithms process all sensor data, then environmental mapping accuracy is improved, but the vehicle's body members are incorrectly added to maps as obstacles
Solution Approach 1:
The patent performs preliminary filtering of point cloud data by removing points within the vehicle's bounding box before the mapping algorithms process the data. This preliminary action prevents false obstacle information from being generated in the first place, ensuring mapping accuracy without requiring post-processing correction of erroneous map data.
Data Source
Figure 1
Figure 2~3
AI summary
Apparatuses, methods, computer programs and systems for environment related data management for a mobile mining vehicle are disclosed. An environment related data management apparatus for a mobile mining vehicle obtains real-time position information of structural body members of the mobile mining vehicle. The apparatus obtains space information of the structural body members of the mobile mining vehicle. The apparatus associates, based on the obtained space information, at least one bounding box with at least one structural body member. The apparatus receives from a sensor configured to scan an environment of the mobile mining vehicle, point cloud data related to the environment of the mobile mining vehicle. In response to determining that data points in the received point cloud data overlap with the associated at least one bounding box, the apparatus generates an overlap indication indicating that at least the overlapping data points are to be excluded.