Role Map Based Image Processing for Overlapping Camera Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video-surveillance systems with overlapping camera fields of view face challenges in efficiently processing images due to high computing power and memory requirements, whether implemented centrally or distributively, leading to increased costs and performance issues.
Innovation Solution
A method and device that generate a role map to optimize image processing by selecting the most appropriate camera source for specific areas of overlap based on occlusion criteria and detection scores, allowing for efficient resource allocation and reduced processing complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If centralized implementation of VCA algorithms is used, then tracking performance is improved, but computing power requirements and memory requirements increase significantly
Solution Approach 1:
The patent segments the video surveillance system into multiple distributed nodes, each capable of independently executing VCA algorithms. Instead of centralizing all processing in one server, the system divides the scene into multiple regions and assigns different processing tasks to different nodes, thereby maintaining high tracking performance while reducing the computing power burden on any single node.
Solution Approach 2:
The patent introduces a spatial dimension to the processing architecture by organizing nodes in a distributed network topology. This allows the system to process video data from multiple cameras simultaneously across different spatial locations, achieving parallel processing that maintains performance while reducing individual node resource requirements.
2Reliability
If centralized implementation of VCA algorithms is used, then tracking performance is improved, but hardware costs increase
Solution Approach 1:
The system segments the hardware infrastructure into multiple standardised nodes rather than requiring one powerful centralized server. Each node uses modest computing resources, and collectively they achieve the performance of a centralized system, thereby reducing overall hardware costs and complexity.
Solution Approach 2:
The patent employs multiple copies of the same VCA algorithm across different nodes in the distributed network. Each node runs an instance of the tracking algorithm, and their results are integrated to achieve comprehensive scene monitoring. This copying approach allows using inexpensive standard hardware instead of expensive specialized equipment.
3Device complexity
If distributed implementation of VCA algorithms is used, then hardware costs are reduced, but tracking performance deteriorates
Solution Approach 1:
The patent merges the results from multiple distributed nodes through a coordination mechanism. Each node processes a portion of the scene independently, and their tracking results are integrated to form a comprehensive view. This merging of distributed processing outcomes maintains tracking performance comparable to centralized systems while using less expensive hardware.
Solution Approach 2:
The distributed nodes exchange feedback information about detected objects and tracking states. This feedback mechanism allows nodes to coordinate their processing, resolve ambiguities, and maintain consistent tracking across the entire scene, thereby preserving tracking performance despite the distributed architecture.
4Device complexity
If distributed implementation of VCA algorithms is used, then resource usage is reduced, but processing efficiency decreases
Solution Approach 1:
The patent segments the processing workload into independent tasks distributed across multiple nodes. Each node processes a specific region or type of detection, avoiding redundant processing. This segmentation maintains high processing efficiency by eliminating duplicate computations while using fewer resources overall compared to centralized processing.
Solution Approach 2:
Each distributed node performs partial processing of the overall scene rather than complete processing. This partial action approach allows parallel execution of multiple processing tasks simultaneously, increasing overall processing efficiency while reducing the resource burden on individual nodes and the system as a whole.
Data Source
AI summary
The present disclosure relates to processing images representing a real scene, the images comprising at least two images obtained from at least two sources of images, a sub-image of each of the at least two images representing an overlapping area of the real scene. A role map comprising a plurality of parts, all the points of each part of the role map being associated with a single identifier of a source of images, is generated during a training phase. To process the images, parts of the role map are selected as a function of an identifier of a considered source of images, and sub-images of an image obtained from the considered source of images are processed, the processed sub-images corresponding to the selected parts of the role map.