Video Target Detection Using Motion Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video surveillance systems face challenges in accurately detecting and tracking individual targets, particularly in crowded scenes and overhead camera views, where groups of targets are often misclassified as a single 'blob' and inanimate objects are not distinguished from legitimate targets.
Innovation Solution
A video processing system that detects moving pixels, identifies line segments or motion blocks, tracks targets, and manages them, enabling the differentiation of multiple objects even when they are closely grouped and accurately identifying human targets in overhead camera views.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional video surveillance algorithms are used, then the system can detect moving objects, but groups of targets are detected as a single blob and individual targets cannot be distinguished
Solution Approach 1:
The patent applies segmentation by dividing the detection process into multiple stages: initial blob detection, followed by segmentation of the blob into individual targets using motion analysis and spatial-temporal constraints. This allows the system to first detect groups as blobs and then separate them into individual targets, resolving the contradiction between detecting moving objects and identifying individual targets.
Solution Approach 2:
The patent introduces temporal dimension by analyzing motion across multiple video frames. By examining the temporal evolution of pixel positions and velocities, the system can distinguish individual targets within a spatial blob, adding a time dimension to the detection process that enables individual target identification without losing information.
2Measurement precision
If conventional human detection algorithms are used, then the system can recognize human beings, but inanimate objects like furniture and shopping carts are not distinguished from legitimate targets
Solution Approach 1:
The patent applies local quality by analyzing local motion patterns and spatial characteristics of different regions within detected blobs. By examining the specific motion dynamics, velocity vectors, and spatial distribution of pixels in different local regions, the system can distinguish between animate targets (humans) and inanimate objects, reducing false alarms while maintaining accurate target classification.
Solution Approach 2:
The patent changes detection parameters dynamically by adjusting motion thresholds, velocity constraints, and spatial-temporal filters based on the specific characteristics of detected objects. This allows the system to adapt to different scenarios and correctly classify targets while filtering out inanimate objects that do not exhibit human-like motion patterns.
3Productivity
If automated video monitoring is implemented, then manual monitoring intensity is reduced, but the system cannot accurately recognize individual targets in crowded scenes
Solution Approach 1:
The patent introduces intermediate processing steps between automatic blob detection and final target identification. Motion analysis modules, velocity calculation units, and spatial-temporal filtering mechanisms serve as intermediaries that process raw detection data and extract meaningful individual target information, enabling automated monitoring to achieve both efficiency and accuracy in crowded scenes.
Data Source
AI summary
A technique for video processing includes: receiving video from a scene; detecting moving pixels in the video; detecting line segments or motion blocks in the video based on the detected moving pixels; identifying targets in the video based on the detected line segments or motion blocks; tracking targets in the video based on the identified targets; and managing the tracked targets in the video.