3D Point Cloud People Classification via LIDAR Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object detection systems, particularly those using thermal cameras or visible light cameras, suffer from high nuisance alarms and performance degradation in challenging environmental conditions, such as weather and lighting variations, leading to misclassification of objects and inefficient motion detection.
Innovation Solution
The use of LIDAR technology to generate three-dimensional point cloud data, which is then segmented to remove background and convert into two-dimensional data for shape classification, enabling resilient object detection and classification regardless of environmental conditions, including occlusion and varying scales.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If thermal cameras or visible light cameras are used for object classification, then classification capability is provided, but nuisance alarm rate increases and performance degrades in challenging environmental conditions
Solution Approach 1:
The patent segments the detection task into multiple stages: initial motion detection by the LIDAR sensor, followed by classification only of detected motion events using thermal/visible cameras. This selective segmentation reduces unnecessary classification of stationary objects, thereby reducing nuisance alarms while maintaining classification capability for actual targets.
Solution Approach 2:
The patent introduces temporal dimension by implementing velocity filtering that analyzes motion over time. By examining velocity vectors and temporal patterns of detected objects, the system distinguishes between stationary background objects and actual moving targets, reducing false alarms caused by static objects being misclassified.
2Adaptability or versatility
If thermal cameras or visible light cameras are used for object detection, then classification is achieved, but detection accuracy decreases in adverse weather and lighting conditions
Solution Approach 1:
The patent uses LIDAR as an intermediary sensor that operates independently of weather and lighting conditions. The LIDAR performs initial detection and provides velocity filtering, enabling the thermal/visible camera system to focus only on verified moving objects, thereby maintaining detection accuracy in adverse conditions.
Solution Approach 2:
The patent replaces reliance on optical mechanisms (thermal/visible cameras) with LIDAR-based detection for initial object identification. Since LIDAR uses laser ranging rather than optical imaging, it is unaffected by weather and lighting, providing reliable detection that compensates for the weaknesses of camera-based systems.
3Duration of action of stationary object
If stationary sensors are used for monitoring, then continuous monitoring is provided, but background objects generate nuisance alarms that confirm motion sensor false alarms
Solution Approach 1:
The patent introduces dynamic velocity filtering to the stationary sensor system. By calculating velocity vectors from sequential LIDAR measurements and filtering based on motion characteristics, the system dynamically distinguishes between stationary background objects and actual moving targets, reducing nuisance alarms while maintaining continuous monitoring capability.
4Adaptability or versatility
If visible camera and thermal camera classifiers are used, then object classification is provided, but motion detection capability is lacking due to technical constraints
Solution Approach 1:
The patent merges LIDAR-based motion detection with camera-based classification into a unified system. The LIDAR handles motion detection and velocity filtering, while the thermal/visible cameras handle classification, creating a complementary hybrid system that overcomes the limitations of using either technology alone.
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
This approach significantly reduces false alarms and improves object detection accuracy, allowing for effective identification of objects of interest like people, even in adverse weather and complex environments, with enhanced capabilities for covert detection and tracking.
Implementation Method 1
receiving, from a sensor, monitoring data associated with the environment, where the monitoring data comprising a series of frames of three-dimensional point cloud information
Data Source
AI summary
A detector may include processing circuitry configured to receive monitoring data associated with an environment being monitored, where the monitoring data includes a series of frames of three-dimensional point cloud information. The processing circuitry may be further configured to perform image segmentation on the monitoring data to generate segmented three-dimensional data associated with the environment, identify one or more objects in the environment based on subtracting background from the segmented three-dimensional data, convert data associated with the one or more objects from the segmented three-dimensional data into two-dimensional object data, and perform object classification on the two-dimensional object data.