Predictive Graph-Map Video Capture for Multi-Camera Surveillance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video surveillance systems often miss detecting fast-moving objects, objects outside the current field of view, or early portions of events due to delayed or inefficient video capture modifications, leading to increased storage costs and reduced effectiveness.
Innovation Solution
A system that uses graph maps to predictively modify video capture operations across multiple networked cameras by sharing video data and alert events, allowing for coordinated adjustments in video capture parameters based on spatial relationships and object detection across camera groups.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If video cameras use trigger conditions to selectively capture high quality video, then storage costs are reduced, but detection delays occur and fast moving objects are missed
Solution Approach 1:
The system performs preliminary actions by having cameras send video capture update messages to neighboring cameras before objects actually enter their fields of view. Graph maps are used to predict object trajectories and proactively notify adjacent cameras to prepare their capture operations, allowing cameras to switch to high-quality mode in advance rather than waiting for trigger conditions, thus eliminating detection delays while maintaining storage efficiency.
Solution Approach 2:
The patent introduces an intermediary communication mechanism where video capture update messages are transmitted between cameras through a network. These messages serve as intermediaries that carry trajectory information and alert events from one camera to neighboring cameras, enabling coordinated capture operations without requiring continuous high-quality recording by all cameras, thus reducing storage costs while improving detection responsiveness.
2Loss of energy
If video cameras operate at lower capture rates to reduce storage usage, then storage costs are reduced, but objects that are more difficult to detect are missed
Solution Approach 1:
The system uses graph maps to predict object trajectories and sends advance notifications to cameras that are about to encounter difficult-to-detect objects. This preliminary action allows receiving cameras to proactively switch to high-quality capture mode before the object enters their field of view, ensuring reliable detection of challenging objects while maintaining low capture rates for cameras not currently monitoring objects of interest, thus preserving storage efficiency.
Solution Approach 2:
The patent implements local quality by allowing each camera to dynamically adjust its capture quality based on its specific situational context. Instead of all cameras operating at uniform high quality, only cameras with predicted objects in their vicinity switch to high-quality mode, while others maintain low capture rates. This localized quality adjustment ensures detection accuracy for objects that are difficult to detect while minimizing overall storage usage.
3Loss of energy
If cameras wait for objects to enter field of view before capturing, then storage is optimized, but early portions of events with critical features are missed
Solution Approach 1:
The system performs preliminary actions by using graph maps to predict object trajectories and sending video capture update messages to neighboring cameras before objects actually enter their fields of view. This advance notification allows cameras to begin high-quality capture operations in advance, ensuring that early portions of events including critical angles, lighting conditions, and other image features are captured, while still optimizing storage by maintaining low capture rates for cameras without predicted objects.
Data Source
AI summary
Systems, video cameras, and methods for predictive adjustment of multi-camera surveillance video data capture based on a network of sub-region graph maps are described. Different groups of networked video cameras are deployed across a region and represented in sub-region graph maps based on the physical locations of the video cameras and interrelated through a master graph map. Each group of networked video cameras includes a master video camera able to connect directly and/or through a control center to the master video cameras in other groups to share video capture update messages in response to regional video alert events.