Multi-Platform Moving Object Detection via Motion Likelihood Accumulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing moving object detection and tracking systems face challenges in maintaining synchronization across multiple platforms, especially when tracking multiple objects over extended periods, and often require human operator interaction, which can delay position updates and render simultaneous tracking impractical.
Innovation Solution
The system accumulates motion likelihoods from sequential image frames to detect moving objects by determining regions with high probability of movement, eliminating image clutter, and associating regions with predicted velocities without relying on third-party targeting or human interaction, using a sensor scan from platforms like UAVs or aircraft.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple platforms are used for moving object detection and tracking, then the coverage area and detection capability are improved, but the synchronization difficulty and system complexity increase
Solution Approach 1:
The patent combines multiple sensor platforms (airborne, spaceborne, ground-based) into a unified detection system that shares common processing infrastructure. The system merges data from different platforms and integrates them with existing radar, infrared, and other sensor systems, creating a consolidated multi-sensor platform that reduces overall system complexity while maintaining enhanced detection capabilities.
Solution Approach 2:
The patent creates a universal processing system that can handle data from multiple different sensor types and platforms simultaneously. The common processing infrastructure is designed to be platform-agnostic, capable of processing imagery from airborne sensors, spaceborne sensors, and ground-based sensors using the same motion likelihood accumulation algorithms, thereby reducing complexity through standardized multi-functional processing.
2Measurement precision
If traditional object detection systems require human operator interaction, then detection accuracy can be improved through expert judgment, but the update rate and productivity decrease
Solution Approach 1:
The patent implements self-service detection algorithms that automatically accumulate motion likelihoods from sequential image frames without requiring human operator intervention. The system autonomously processes imagery, identifies moving objects through motion likelihood accumulation, and generates detection results, thereby maintaining high detection accuracy while achieving rapid update rates suitable for real-time tracking of multiple objects.
Solution Approach 2:
The patent replaces the mechanical human-in-the-loop detection process with automated computational algorithms. Instead of relying on human operators to analyze imagery and identify objects, the system uses motion likelihood accumulation algorithms that automatically detect moving objects by comparing sequential frames, substituting human cognitive processing with automated image processing while improving both speed and consistency.
3Measurement precision
If precise timing synchronization is maintained across multiple platforms, then tracking accuracy is improved, but the operational difficulty and time loss increase
Solution Approach 1:
The patent implements preliminary synchronization procedures that establish timing relationships between multiple platforms before detection operations begin. The system pre-configures synchronization parameters and establishes reference time bases for all platforms, allowing subsequent detection operations to proceed with minimal real-time synchronization adjustments, thereby maintaining high tracking accuracy while reducing the time spent on synchronization maintenance during extended surveillance.
4Adaptability or versatility
If the system tracks multiple moving objects simultaneously, then the surveillance coverage is improved, but the complexity of maintaining position updates increases
Solution Approach 1:
The patent segments the tracking process into independent object-specific processing streams, where each detected moving object is assigned its own detection and tracking pipeline. The motion likelihood accumulation and position update algorithms operate independently for each object, allowing the system to scale to multiple simultaneous tracks without proportionally increasing overall system complexity. Each object's position updates are processed separately through standardized algorithms.
Data Source
AI summary
Moving object detection methods and systems are described. In an embodiment, motion likelihoods are accumulated from sets of sequential image frames which are generated from a sensor scan of one or more moving objects. Regions that each indicates a probable moving object from the accumulated motion likelihoods are determined, and the one or more moving objects are then detected from the respective regions that indicate a probable moving object.