Thermal Target Extraction Using Histogram Differentiation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting and extracting warm or hot objects from thermal images are prone to false results due to inadequate segmentation techniques, especially when dealing with camera movements and scene changes, and are not power-efficient for handheld devices.
Innovation Solution
A method that estimates ambient temperature, differentiates histograms to set a reliable threshold, and filters dynamic threshold values using a weighted average to accurately segment targets from background clutter in thermal video frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If clustering or histogram methods are used for segmentation, then power consumption is reduced, but false target classification increases
Solution Approach 1:
The patent segments the thermal image into multiple clusters representing different temperature ranges (e.g., background, mid-temperature objects, hot targets). By dividing the temperature spectrum into distinct segments and analyzing the distribution of pixels in each segment, the method achieves both low power consumption (through simple histogram-based clustering) and high reliability (by examining multiple segments rather than relying on a single threshold).
Solution Approach 2:
Instead of using a single threshold value that may cause false classifications, the patent applies multiple threshold levels corresponding to different temperature clusters. This partial action approach examines pixel distributions across several temperature bands, allowing the system to identify true targets by their characteristic presence in specific clusters while rejecting false alarms that don't match the expected cluster pattern.
2Power
If clustering or histogram methods are used for segmentation, then computational power is reduced, but ability to account for scene changes deteriorates
Solution Approach 1:
The patent dynamically adapts the clustering parameters and threshold values based on the current scene characteristics. By continuously analyzing the pixel distribution across temperature clusters and adjusting the segmentation parameters accordingly, the system maintains adaptability to scene changes (such as camera movements or varying environmental conditions) while keeping computational requirements low through efficient histogram-based methods.
3Measurement precision
If Fourier analysis, pattern classification, or template matching are used, then segmentation accuracy is improved, but power consumption increases
Solution Approach 1:
The patent replaces complex, computationally expensive segmentation algorithms (Fourier analysis, pattern classification, template matching) with a simpler, lower-power alternative based on histogram analysis and cluster-based thresholding. While the simpler method requires less power, it achieves comparable segmentation accuracy by focusing on the statistical distribution of pixel temperatures rather than attempting complex pattern recognition, effectively using a 'cheaper' computational approach that consumes less energy.
Data Source
AI summary
A method for extracting a target from a series of images includes the steps of: (a) estimating an ambient temperature value of pixels in the series of images; (b) finding a band of pixel values having temperature values above the ambient temperature value, the band of pixel values forming a histogram; and (c) differentiating the histogram to estimate a threshold. Also included are steps (d) extracting the target having pixel values above the threshold; and (e) colorizing the target for display.