Stationary Target Detection via Dual Background Model Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video processing systems face challenges in accurately detecting and classifying stationary targets in crowded or high-traffic areas due to frequent occlusions and difficulties in distinguishing new motion from original stationary target motion.
Innovation Solution
The system constructs multiple background models with different update parameters to differentiate between stationary and moving targets, using a change detecting engine to compare pixel differences and a blob classifying engine to determine stationary targets, thereby generating alerts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single background model is used for stationary target detection, then the system is simple to implement, but detection accuracy deteriorates in crowded or high-traffic areas due to frequent occlusions and inability to distinguish new motion from original stationary target motion
Solution Approach 1:
The patent divides the background model construction into multiple parallel models (first background model and second background model) with different update parameters. This segmentation allows each model to capture different temporal characteristics of the scene, improving stationary target detection accuracy by comparing differences between models while maintaining manageable complexity through modular architecture.
Solution Approach 2:
The patent changes the update parameters (learning rates) of different background models to create diverse temporal representations. The first background model uses a first update parameter while the second uses a second update parameter, allowing the system to detect stationary targets by identifying pixels that differ between models, thus resolving the contradiction between detection accuracy and system complexity.
2Reliability
If background subtraction is used to detect moving targets, then moving targets can be detected and tracked, but stationary target detection becomes problematic when other targets move between the stationary target and camera
Solution Approach 1:
The patent segments the detection process into two parallel background models with different temporal responses. By comparing pixel differences between these segmented models, the system can reliably identify stationary targets even when occlusions occur, as the dual-model approach provides redundant information that helps distinguish true stationary objects from transient occlusions.
Solution Approach 2:
The patent introduces a difference detection mechanism as an intermediary between the two background models and the final stationary target identification. This intermediary process compares pixel values from both models and isolates true stationary targets by filtering out transient motions, thereby improving reliability while managing detection complexity.
Data Source
AI summary
A sequence of video frames of an area of interest is obtained. A first background model of the area of interest is constructed based on a first parameter. A second background model of the area of interest is constructed based on a second parameter, the second parameter being different from the first parameter. A difference between the first and second background models is determined. A stationary target is determined based on the determined difference. An alert concerning the stationary target is generated.