PoE Switch Image Evaluation for Network Load Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Surveillance systems with many cameras connected to a switch become overloaded, leading to network paralysis and limitations in expansion, due to high data transmission from video streams, which also affects internal communication and alarm messaging.
Innovation Solution
Implementing an image evaluation device at each camera connected to the switch, performing initial data analysis, and transmitting only the results to the monitoring center, using inexpensive standard cameras and reducing the load on the network and monitoring center, with additional features like ring memory for storing images and a communication computer for joint evaluations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If many cameras are connected to a switch and transmit video streams simultaneously, then the surveillance coverage is expanded, but the network becomes overloaded and paralyzed
Solution Approach 1:
The switch performs preliminary image evaluation and detects movement in video frames before transmitting data to the surveillance center. This preliminary action filters out static scenes, so only frames with detected movement are transmitted, dramatically reducing network traffic while maintaining surveillance coverage
Solution Approach 2:
The invention extracts only the essential information (movement detection results) from the full video streams at the switch level. Instead of transmitting complete video data from all cameras, only the relevant movement detection data is extracted and sent to the surveillance center, reducing network load while preserving surveillance effectiveness
2Loss of substance
If image evaluation is performed in cameras, then data transmission volume is reduced, but processor performance requirements and system cost increase
Solution Approach 1:
The switch acts as an intermediary device between the cameras and the surveillance center. Instead of requiring complex image evaluation in the cameras themselves, the switch performs the movement detection function, serving as a mediator that processes video data from multiple cameras and transmits only relevant information to the surveillance center
Solution Approach 2:
The switch utilizes its existing hardware resources (processor, memory) to perform image evaluation functions. Rather than adding dedicated image evaluation hardware to each camera, the switch serves itself by handling the computational tasks using its available capabilities, thereby reducing overall system complexity
3Loss of substance
If compression methods are used to reduce data volume, then network load is reduced, but image analysis quality deteriorates
Solution Approach 1:
The switch performs preliminary movement detection on uncompressed or minimally compressed video frames before transmission. By detecting movement in the original quality frames and then transmitting only the movement detection results or selectively transmitted frames, the system maintains image analysis quality while reducing overall data transmission volume
4Extent of automation
If the monitoring center performs all image evaluation, then centralized control is maintained, but the monitoring center becomes overloaded
Solution Approach 1:
The image evaluation function is segmented between the switch and the surveillance center. The switch handles the initial movement detection for multiple cameras, filtering out static scenes. The surveillance center then receives only the reduced set of movement-related data, allowing centralized control to be maintained while distributing the processing load and preventing overload at the monitoring center
Data Source
Figure 1
AI summary
The surveillance system has at least one group of cameras, each connected via an Ethernet cable and a PoE power supply (2) to a switch (8), which in turn is connected to a monitoring station via an Ethernet cable (10). To prevent overloading the Ethernet cable (10) to the monitoring station, the switch (8) is, according to the invention, equipped with an image processing unit (4) for each camera, which is connected to the input signal lines of the cameras in the group. Data frames representing the results of the image processing unit (4) are transmitted via the switch (8) to the monitoring station. This allows the use of conventional PoE cameras.In order to be able to retrieve all the information in the last seconds before an alarm, in addition to the image evaluation unit (4) for each camera, a ring buffer can be provided in the switch (8) to store the images of the last seconds of each camera.