Obstacle Detection Filtering Pedestrian-Carried Objects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing obstacle detection methods generate false warnings when pedestrians pass through monitored regions, leading to low accuracy in identifying actual obstacles.
Innovation Solution
Perform object and pedestrian detection on monitored images, associating object identification boxes with pedestrian boxes to filter out objects carried by pedestrians, and issue alarms only for non-associated boxes, using template images to account for lighting variations and multi-thread processing to handle data efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If object detection is performed on all detected objects in the monitored region, then detection coverage is improved, but false alarms increase due to pedestrians being misidentified as obstacles
Solution Approach 1:
The detection process is segmented into two independent detection tasks: object detection and pedestrian detection. By separating the detection of general objects from the detection of pedestrians, the system can independently process and analyze each type of target, thereby avoiding misidentification of pedestrians as obstacles while maintaining comprehensive detection coverage.
Solution Approach 2:
The system extracts and removes pedestrian identification boxes from the object identification group. By taking out the pedestrian detection results and excluding them from the obstacle detection set, the system eliminates false alarms caused by pedestrians being misclassified as obstacles, while preserving the detection of actual obstacles.
2Reliability
If pedestrian detection is added to the detection system, then false alarm reduction is achieved, but system complexity increases
Solution Approach 1:
The object detection model and pedestrian detection model are merged into a unified detection framework that processes images simultaneously for both object and pedestrian identification. This combined approach shares computational resources and processing pipelines, reducing the overall system complexity compared to running separate detection systems while maintaining the ability to distinguish pedestrians from obstacles.
Solution Approach 2:
The detection system is designed with multi-functionality, where the detection models can identify multiple types of targets (objects and pedestrians) using a unified processing architecture. This universal detection capability allows the system to perform both object detection and pedestrian detection through integrated algorithms, simplifying the system structure while achieving reliable obstacle detection.
3Reliability
If association processing between object boxes and pedestrian boxes is performed, then false alarm reduction is improved, but processing time increases
Solution Approach 1:
The system performs preliminary detection of both objects and pedestrians in parallel before conducting association processing. By pre-identifying all potential targets and their bounding boxes beforehand, the subsequent association step only needs to perform simple coordinate matching and comparison operations, significantly reducing the processing time required for the association step while maintaining accurate false alarm reduction.
Solution Approach 2:
The system uses bounding box coordinates as simplified representations (copies) of the actual objects and pedestrians. Instead of performing complex object recognition and comparison during association processing, the system works with copied spatial information (coordinates and dimensions of detection boxes), which greatly accelerates the association process while preserving the essential information needed to distinguish pedestrians from obstacles.
Data Source
AI summary
A method for detecting obstacles includes: acquiring a monitored image of a monitored region, performing object detection on the monitored image, performing pedestrian detection on the monitored image, in response to acquiring an object identification group comprising at least one object identification box by the object detection, and acquiring at least one pedestrian identification box by the pedestrian detection, acquiring a to-be-determined object identification group by removing an object identification box in the object identification group associated with the at least one pedestrian identification box, wherein the object identification box comprises an image of an object in the monitored image, and the pedestrian identification box comprises an image of a pedestrian in the monitored image, and in response to determining that an object identification box is present in the to-be-determined object identification group, determining that an obstacle is present in the monitored region.