Network Camera Object Detection Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems with network cameras face challenges in real-time object detection due to low throughput and limited hardware resources, leading to inaccurate detection and increased load on servers as the number of connected cameras increases.
Innovation Solution
An image processing system where network cameras capture images, detect likelihoods, and transmit tentative object information to an image processing apparatus for further detection, distributing the processing load and improving accuracy by using a larger recognition dictionary for redetection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If object detection processing is performed on low-throughput cameras with limited hardware resources, then detection functionality is provided, but detection accuracy deteriorates due to limited recognition dictionary capacity
Solution Approach 1:
The detection processing is segmented between the camera side and server side. The camera performs preliminary detection with a small recognition dictionary and transmits tentative results to the server, which performs detailed detection with a large recognition dictionary. This segmentation allows each component to operate within its resource constraints while achieving overall high accuracy.
Solution Approach 2:
The network transmission of tentative object information acts as an intermediary mechanism. Instead of requiring the camera to perform all detection processing, the system uses network communication to transfer intermediate results to the server for final processing, enabling the camera to maintain simplicity while the server provides high accuracy.
2Adaptability or versatility
If the number of connected cameras increases, then system coverage is improved, but server processing load increases excessively
Solution Approach 1:
The processing load is segmented and distributed across multiple cameras rather than concentrated on a single server. Each camera performs local detection processing independently, so adding more cameras distributes the detection workload across more devices rather than increasing server load proportionally.
Solution Approach 2:
Each camera performs self-service detection processing using its own recognition dictionary and processing units. This autonomous capability reduces the burden on the server, as cameras handle detection tasks independently before only transmitting necessary results to the server for coordination and detailed processing.
3Measurement precision
If high-resolution images are processed for detailed object detection, then detection accuracy is improved, but processing time increases
Solution Approach 1:
The processing is segmented into two stages: preliminary processing at the camera level using lower computational requirements, and detailed processing at the server level for high-resolution analysis. This segmentation allows the system to maintain speed during initial detection while providing high accuracy through subsequent detailed processing only when necessary.
Solution Approach 2:
The camera performs partial detection action sufficient for preliminary identification, transmitting only necessary information to the server. The server then performs excessive or detailed processing only for objects that require high-resolution analysis, avoiding unnecessary processing time for all images while maintaining accuracy where needed.
Data Source
AI summary
There is provided an image processing system in which an image capture apparatus and an image processing apparatus are connected to each other via a network. When a likelihood indicating the probability that a detection target object detected from a captured image is a predetermined type of object does not meet a designated criterion, the image capture apparatus generates tentative object information for the detection target object, and transmits it to the image processing apparatus. The image processing apparatus detects, from detection targets designated by the tentative object information, a detection target as the predetermined type of object.