LiDAR Intrusion Detection That Excludes the Working Machine
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing intrusion detection systems fail to accurately distinguish a work machine at work in a work area from other objects entering the area, leading to erroneous detections.
Innovation Solution
An intrusion detection system that utilizes a machine position acquisition unit, point cloud data acquisition, position calculation, and a specifying unit to identify the work machine's position and shape, allowing accurate detection of other objects entering the work area by distinguishing the work machine's portion in point cloud data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If simple object detection is used, then detection speed is improved, but detection accuracy deteriorates due to inability to distinguish work machine from intruding objects
Solution Approach 1:
The patent segments the detection process into multiple stages: first detecting objects using simple methods, then identifying work machines using position information, and finally distinguishing intruding objects by comparing positions. This segmentation allows the system to maintain fast initial detection while achieving high accuracy through subsequent filtering.
Solution Approach 2:
The patent extracts position information as a key feature to distinguish work machines from intruding objects. By extracting and utilizing position data from the point cloud, the system can filter out work machines from detection results, thereby improving accuracy without sacrificing detection speed.
2Measurement precision
If work machine position information is integrated into detection, then detection accuracy is improved, but system complexity increases
Solution Approach 1:
The patent makes the detection system multi-functional by integrating multiple capabilities: object detection, work machine identification using position information, and intrusion determination. This universal approach allows a single system to handle various tasks without requiring separate specialized systems, thereby managing complexity through functional integration.
Solution Approach 2:
The patent creates a simplified representation (copy) of the work machine's position and spatial characteristics from the point cloud data. This copy is then used for comparison and identification purposes, reducing the complexity of processing full 3D models while maintaining detection accuracy.
3Measurement precision
If point cloud data processing is performed, then detection accuracy is improved, but computational load increases
Solution Approach 1:
The patent extracts only the necessary position information from the point cloud data for work machine identification, rather than processing the entire point cloud in detail. This selective extraction significantly reduces computational load while maintaining the accuracy needed to distinguish work machines from intruding objects.
Solution Approach 2:
The patent applies partial processing to the point cloud data, focusing computational resources only on regions or features relevant to work machine identification. By avoiding complete processing of all data, the system achieves sufficient accuracy with reduced computational energy consumption.
Data Source
AI summary
The intrusion detection system includes a GNSS sensor that acquires a position of a work machine at work in a work area, a LiDAR that is disposed outside the work area and acquires point cloud data indicating a distance up to an object located inside or outside the work area, a position calculation unit that calculates positions of respective points of the point cloud data, a specifying unit that specifies a portion corresponding to the work machine from the point cloud data as specific data, based on the position of the work machine and the positions of the respective points of the point cloud data, and a detection unit that detects that an object has entered the work area, based on the point cloud data excluding the specific data.