Abandoned Object Detection False Alarm Reduction via Background Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing abandoned object detection systems in video surveillance face challenges with false detection due to environmental changes, particularly in complex scenes like continuous light changes, leading to inaccurate identification of abandoned objects.
Innovation Solution
A recognition method and apparatus that perform foreground detection on current frame images, cluster contours into candidate regions, and compare these regions with a reference background image to determine if they match, thereby identifying and correcting false detections by feeding back results to the foreground detection module.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If background modules are updated according to historical information of each pixel to detect abandoned objects, then abandoned object detection capability is improved, but false detection increases due to environmental changes such as continuous light variations
Solution Approach 1:
The patent segments the detection process into multiple stages: foreground detection to obtain candidate regions, contour extraction to identify object boundaries, and matching verification against background database. This multi-stage segmentation allows the system to filter out false detections caused by environmental changes while maintaining sensitivity to real abandoned objects.
Solution Approach 2:
The patent introduces an intermediary matching verification mechanism that compares extracted contours against a background database before final detection. This intermediary step acts as a mediator between foreground detection and final judgment, eliminating false alarms caused by light variations while preserving true abandoned object detections.
2Difficulty of detecting and measuring
If foreground detection is performed on current frame images to identify abandoned objects, then detection sensitivity is improved, but false alarms increase in complex scenes with environmental changes
Solution Approach 1:
The patent performs preliminary actions by maintaining a background database that stores historical background information before detection. This preliminary preparation allows the system to quickly compare current frames against known background patterns, enabling high detection sensitivity while filtering out false alarms from environmental changes through pre-established reference data.
Solution Approach 2:
The patent implements a feedback mechanism where detection results are verified by comparing against the background database. This feedback loop allows the system to adjust and refine detections in real-time, confirming true abandoned objects while eliminating false alarms caused by continuous light variations and other environmental factors.
Data Source
AI summary
Embodiments of this disclosure provide a recognition method and apparatus for false detection of an abandoned object and an image processing device. The method includes: performing foreground detection on a current frame image, so as to obtain a foreground image of the current frame image, the foreground image comprising an abandoned object foreground image; clustering contours in the abandoned object foreground image into a group, and take the group as a candidate region of an abandoned object; and comparing the candidate region of the abandoned object with a reference background image, judging whether the candidate region of the abandoned object is matched with the reference background image, and a match, deeming that the abandoned object is false detection.