Moving Object Detection Using Adaptive Frame Interval Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional moving object detection techniques in blind-spot camera systems are insufficiently accurate for detecting objects at distant spots due to unstable calculations and the need for sub-pixel estimation, which can lead to erroneous judgments and increased memory requirements, especially when detecting small movements.
Innovation Solution
A moving object detection apparatus that calculates the moving distance of characteristic points between successive image frames and uses an adaptive time difference control to track movement, allowing for the detection of both long and short distance movements without requiring information on the distance of image areas, and employs a block matching method with a Sum of Absolute Difference evaluation to reduce calculation complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a gradient method with sub-pixel estimation is used to detect distant moving objects, then measurement precision is improved, but reliability deteriorates due to unstable calculations
Solution Approach 1:
The patent segments the moving object detection into two distinct phases: first detecting large movements using block matching method, then detecting small movements using gradient method with sub-pixel estimation. This segmentation allows each method to operate in its optimal range, improving overall reliability while maintaining precision for distant objects.
Solution Approach 2:
The patent dynamically switches between detection methods based on the magnitude of movement detected. When large movement is detected, block matching is used; when small movement is detected, gradient method with sub-pixel estimation is applied. This dynamic adaptation resolves the contradiction by selecting the appropriate method for each specific detection scenario.
2Productivity
If the time interval between frames is expanded to detect high-speed moving objects, then productivity is improved, but measurement precision deteriorates due to out-of-search-scope points
Solution Approach 1:
The patent dynamically adjusts the frame interval based on detection needs. For high-speed objects, larger time intervals are used to capture significant movement; for distant slow-moving objects, smaller intervals with sub-pixel estimation are used. This dynamic adjustment maintains precision across different object types and speeds.
Solution Approach 2:
The detection process is segmented into multiple passes with different time intervals. First, large movements are detected with longer intervals; then, remaining small movements are detected with shorter intervals and sub-pixel estimation. This segmented approach resolves the contradiction by applying appropriate time intervals to different detection stages.
3Device complexity
If block matching method is used for sub-pixel flow calculation, then device complexity is reduced, but measurement precision deteriorates due to zero flow calculation
Solution Approach 1:
The patent segments the flow calculation into two stages: first using block matching method for integer-pixel accuracy with simple processing, then using gradient method with sub-pixel estimation for fractional-pixel accuracy when needed. This segmentation maintains simplicity for most cases while achieving high precision when required.
Solution Approach 2:
The patent dynamically selects the calculation precision level based on detection requirements. For most objects, integer-pixel block matching is used for simplicity; for distant objects requiring higher precision, sub-pixel gradient method is applied. This dynamic selection resolves the contradiction between complexity and precision.
4Productivity
If more image frames are stored for expanded intervals, then productivity is improved, but loss of substance increases due to greater memory requirements
Solution Approach 1:
The patent segments frame storage requirements by detecting and processing large movements first, which reduces the number of frames that need to be retained for subsequent processing. This segmentation significantly reduces memory requirements while maintaining the ability to detect high-speed objects.
Solution Approach 2:
The patent performs preliminary detection of large movements before requiring extensive frame storage. By identifying and processing significant movements early, the system reduces the memory burden for subsequent detailed analysis, resolving the contradiction between productivity and memory usage.
Data Source
AI summary
The approaching object detection unit in a moving object detection apparatus for a moving picture calculates a moving distance of each characteristic point in an image frame obtained at time point t−1, on the basis of an image frame obtained at time point t and an image frame obtained at time point t−1, and on the basis of the image frame obtained at time point t−1 and an image frame obtained at time point t+m, a moving distance of a characteristic point is in the image frame obtained at time point t−1 and has a moving distance to be less than a prescribed value.