IR Object Counting via Reflectance Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing infrared (IR) imaging systems face challenges in accurately determining the number of objects in an image, especially when objects are moving and illumination is inadequate, which is common in real-world applications such as vehicle occupancy detection and security surveillance.
Innovation Solution
A novel system and method that uses an IR imaging system to collect intensity values for each pixel, classify them based on correlation coefficients or best fitting reflectances, and separate objects from the background to determine the total number of objects in the IR image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If face recognition systems use traditional camera illumination requirements, then recognition accuracy is improved, but system reliability deteriorates under inadequate illumination conditions
Solution Approach 1:
The system changes the wavelength parameter of illumination by using infrared illumination instead of visible light, allowing operation under conditions where traditional illumination fails. The infrared camera captures reflected infrared light from facial features, enabling recognition when visible light illumination is inadequate.
Solution Approach 2:
The system replaces the mechanical/optical requirement for visible light illumination with an infrared-based detection system. Instead of relying on visible light cameras that need adequate illumination, the system uses infrared cameras that can detect thermal radiation and reflected infrared light, substituting the illumination dependency with thermal/infrared detection capabilities.
2Speed
If the subject is moving while passing in front of the camera, then detection speed is improved, but measurement precision deteriorates due to motion blur and positioning challenges
Solution Approach 1:
The system uses periodic infrared illumination pulses synchronized with the camera shutter to capture multiple frames of the moving subject. By illuminating periodically and capturing at specific phases, the system can detect moving subjects more reliably and reduce motion blur effects while maintaining detection speed.
3Difficulty of detecting and measuring
If background subtraction methods are used to separate objects, then object detection is improved, but reliability deteriorates when background changes or objects are stationary
Solution Approach 1:
The system changes from spatial/temporal background subtraction to spectral parameter analysis. By analyzing the spectral signature of different materials at multiple wavelengths, the system can distinguish objects from backgrounds based on their unique reflectance characteristics, regardless of whether the background is changing or the object is stationary.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Effectively separates living objects from the background, enabling accurate counting of occupants in vehicles or other scenarios, even under challenging conditions, with robustness demonstrated through correlation coefficient calculations and reflectance-based classification methods.
Implementation Method 1
using an IR imaging system, a total of N intensity values are collected for each pixel in an IR image
Implementation Method 2
intensity values are calculated using reflectances which have been estimated for a plurality of known materials such as, for example, hair and skin
Data Source
AI summary
What is disclosed is a novel system and method for determining the number of objects in an IR image obtained using an IR imaging system. In one embodiment, a total of N intensity values are collected for each pixel in an IR image using a IR imaging system comprising an IR detection device and an IR Illuminator. Intensity values are retrieved from a database which have been estimated for a plurality of known materials, such as skin and hair. A classification is determined for each pixel in the IR image using either a best fitting method of a reflectance, or a correlation method. Upon classification, a total number of objects in the IR image can be determined. The present system and method finds its intended uses in of real world applications such as, determining the number of occupants in a vehicle traveling in a HOV/HOT lane.