Surveillance Image Selection for Target Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Operators face significant challenges in efficiently identifying targets from vast amounts of recorded video due to small or improperly oriented images produced by intrusion detection systems, leading to increased personnel costs and workload.
Innovation Solution
A representative image generation device that identifies and displays images of objects with enhanced visual recognizability by tracking and calculating suitability scores based on factors like object size, aspect ratio, direction, human likeness, and luminance, ensuring the selected image provides better visual recognition than a predetermined threshold.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If intrusion detection functions are used to detect persons or vehicles, then the search time is reduced, but the detected objects appear small or in inappropriate orientations making identification difficult
Solution Approach 1:
The system performs preliminary actions by tracking the object through multiple frames and pre-calculating suitability scores for each frame before selection. This allows the system to identify the optimal frame for identification in advance, ensuring both time efficiency and high visual recognizability without requiring operators to manually review multiple frames.
Solution Approach 2:
The system creates a representative image (copy) from the original surveillance footage by selecting and processing the most suitable frame. This representative image serves as an optimized copy that preserves essential identification features while eliminating the time-consuming need to view the entire video sequence, thus resolving the contradiction between speed and recognition quality.
2Measurement precision
If operators visually search vast amounts of recorded video one by one, then accurate identification is possible, but personnel costs and workload increase significantly
Solution Approach 1:
The system performs self-service by automatically selecting and preparing the representative image without requiring operator intervention for frame-by-frame review. The automated suitability score calculation and frame selection process enables the system to identify targets accurately while dramatically reducing operator workload and increasing productivity simultaneously.
Solution Approach 2:
The system replaces the mechanical process of manual video review with an automated computational approach. By substituting operator visual inspection with algorithmic suitability score calculation and automatic frame selection, the system maintains high identification accuracy while eliminating the productivity loss associated with manual searching of vast video volumes.
3Adaptability or versatility
If the most balanced front face image is selected from sequential images, then objective selection is achieved, but the image may not be optimal for distinguishing the target object in surveillance contexts
Solution Approach 1:
The system applies local quality by adjusting image selection criteria based on the specific surveillance context rather than using a universal front-face detection approach. The suitability score calculation incorporates local factors such as object size, position in frame, and temporal consistency, which are specifically optimized for surveillance target identification rather than general face detection, thereby improving both adaptability and reliability.
Solution Approach 2:
The system changes parameters by transitioning from static front-face detection criteria to dynamic suitability scores that evaluate multiple frames over time. This parameter change allows the system to select images based on comprehensive criteria including object size, frame position, temporal stability, and detection confidence, significantly improving target identification reliability while maintaining flexibility across different surveillance scenarios.
Data Source
AI summary
A representative image generation device includes circuitry that identifies a selected image from a plurality of recorded images of an object. The plurality of recorded images are captured as the object traverses a predetermined area. The device prepares the selected image for display on a display device, wherein the selected image provides a visual recognizability of the object greater than a predetermined threshold.