Moving Object Detection Using Segmented Motion Vector and Difference Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing moving object detection systems face challenges in detecting fast-moving objects due to object blur, which prevents the calculation of motion vectors, and struggle with distinguishing between camera motion and object motion, especially in unfixed cameras, leading to reduced detection accuracy.
Innovation Solution
A moving object detection apparatus that combines motion vector calculation and inter-frame difference analysis, using multiple detection devices to integrate results and determine object presence, with separate time intervals for each method to balance real-time performance and accuracy, and includes features for panning determination and focus adjustment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion vector calculation is used for moving object detection, then detection accuracy for slow-moving objects is improved, but detection capability for fast-moving objects deteriorates due to object blur
Solution Approach 1:
The detection system is segmented into two independent detection devices: a first moving object detection device that uses motion vector calculation for slow-moving objects, and a second moving object detection device that uses inter-frame difference for fast-moving objects. Each device handles specific speed ranges, allowing both slow and fast moving objects to be detected accurately without interference between methods.
Solution Approach 2:
The patent merges the outputs of two different detection methods into a unified detection result. The integration device combines detection results from both the motion vector-based first device and the inter-frame difference-based second device, allowing the system to detect both slow-moving objects (where motion vectors are reliable) and fast-moving objects (where inter-frame difference is necessary).
2Measurement precision
If motion vector calculation is performed, then detection accuracy for slow-moving objects is improved, but calculation cost and processing time increase
Solution Approach 1:
The processing workload is segmented and distributed to two different detection devices with different computational requirements. The first device performs computationally intensive motion vector calculation only where needed (for slow-moving objects), while the second device performs simpler inter-frame difference operations for fast-moving objects, optimizing overall system efficiency.
Solution Approach 2:
The system dynamically selects which detection method to use based on the characteristics of the detected objects. By continuously monitoring and comparing detection results, the system can switch between motion vector calculation and inter-frame difference methods depending on object speed and movement characteristics, optimizing real-time performance.
3Adaptability or versatility
If inter-frame difference is used for moving object detection, then detection capability for fast-moving objects is improved, but ability to distinguish camera motion from object motion deteriorates
Solution Approach 1:
The detection system is segmented into two independent detection devices: a first moving object detection device that uses motion vector calculation for slow-moving objects, and a second moving object detection device that uses inter-frame difference for fast-moving objects. Each device handles specific speed ranges, allowing both slow and fast moving objects to be detected accurately without interference between methods.
Solution Approach 2:
The integration device acts as an intermediary that receives detection results from both detection methods and mediates between them. It compares and reconciles the outputs of motion vector-based and inter-frame difference-based detection, using contextual information to distinguish between camera motion and object motion, thereby maintaining detection accuracy while preserving the ability to detect fast-moving objects.
Data Source
AI summary
A moving object detection apparatus comprising: an image acquisition device; a first moving object detection device; a difference image generation device; a second moving object detection device for detecting existence/nonexistence of the moving object based on the difference image generated by the difference image generation device; and an integration device for integrating a detection result by the first moving object detection device and a detection result by the second moving object detection device and determining that the moving object is detected in a case where the moving object is not detected by at least the first moving object detection device and the moving object is detected by the second moving object detection device.