Convolutional Neural Network Object Classification Using Visible and Invisible Light Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current surveillance systems rely on human operators for object detection and classification in images, which is inefficient and lacks real-time capability, especially in conditions like fog or low light where visible light images are obstructed, limiting the system's ability to accurately identify objects.
Innovation Solution
A method using a convolutional neural network that processes both visible and invisible light images, where the network is trained on pairs of images, one taken with visible light and the other with invisible light, to classify and detect objects of interest, enabling the system to identify objects even when they are obscured in visible light images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If human operators are used for object detection and classification, then the system can identify objects in visible light images, but the process is inefficient and lacks real-time capability
Solution Approach 1:
The patent replaces the mechanical system of human visual inspection with an automated electronic system consisting of image capture devices and processing circuits that automatically detect and classify objects in real-time, eliminating the time loss associated with manual review while maintaining or improving detection accuracy
Solution Approach 2:
The system performs self-service by automatically analyzing captured images through processing circuits that identify objects and generate alerts without requiring human intervention, enabling continuous real-time operation and eliminating the bottleneck of manual detection
2Reliability
If only visible light images are used, then the system can capture clear images in normal conditions, but objects become obscured in fog or low light conditions
Solution Approach 1:
The patent merges multiple image capture devices with different operational characteristics (visible light and invisible light sensors) into a single surveillance system, allowing the system to switch between or combine data from both sources to maintain reliable object identification across varying environmental conditions
Solution Approach 2:
The surveillance system achieves multi-functionality by incorporating both visible light imaging for normal conditions and invisible light imaging for challenging conditions like fog or low light, making the system adaptable to a wider range of environments while maintaining consistent detection reliability
3Productivity
If the system uses automated processing, then real-time detection is achieved, but the complexity of the system increases
Solution Approach 1:
The patent segments the automated detection system into distinct functional modules including image capture devices, processing circuits, and alert generation components, allowing each segment to be optimized independently while working together to achieve real-time detection without overwhelming system complexity
Data Source
AI summary
Methods, systems, and techniques for classifying and/or detecting objects using visible and invisible light images. A visible light image and an invisible light image are received at a convolutional neural network (CNN). The visible light image depicts a region-of-interest imaged using visible light. The invisible light image depicts at least a portion of the region-of-interest imaged using invisible light, and at least one of the images depicts an object-of-interest within the portion of the region-of-interest shared between the images. The CNN then classifies and/or detects the object-of-interest using the images. The CNN may be trained to perform this classification and/or detection using pairs of visible and invisible light training images.