Selective Colorization of Thermal Images via Histogram Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional grayscale and colorization techniques for thermal and low-light images fail to effectively highlight objects of interest, leading to cluttered and difficult-to-analyze images, especially in low-contrast scenes, where objects may blend with the background, and combining these images can result in lost fine detail.

Innovation Solution

Selective colorization of thermal images based on intensity levels, using histogram analysis to determine threshold pixel ranges and applying color maps to emphasize hotter or colder objects, and computational combination of thermal and low-light images to enhance edge detection and highlight differences.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If traditional grayscale or full colorization techniques are applied to thermal images, then the entire image is processed uniformly, but objects of interest are not effectively highlighted and the image remains cluttered

Engineering Contradiction:
Improveimage processingVSAvoidobject identification
Core Design Contradiction:
Ease of manufactureVSDifficulty of detecting and measuring

Solution Approach 1:

The patent applies selective colorization by determining a subset of pixels to colorize based on intensity thresholds and histogram analysis, rather than applying colorization uniformly to the entire image. This local quality approach highlights specific regions of interest (such as hotter or colder objects) while leaving other areas in grayscale, thereby reducing clutter and improving object identification.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent segments the image processing task into distinct steps: generating a histogram, determining intensity thresholds, identifying pixels that meet the thresholds, and applying colorization only to those segmented subsets of pixels. This segmentation allows targeted enhancement of specific features without processing the entire image uniformly.

Inventive Principle:
Principle #1Segmentation

2Quantity of substance

If thermal and low-light images are combined to enhance visibility, then more information is available, but fine detail is lost in the combination

Engineering Contradiction:
Improveinformation contentVSAvoidfine detail
Core Design Contradiction:
Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent applies selective colorization to the combined thermal and low-light images, focusing computational resources on specific pixel subsets that meet intensity criteria. This localized processing preserves fine details in regions where objects of interest are detected, while avoiding the loss of detail that would occur if the entire combined image were processed uniformly or if detailed processing were applied to all regions.

Inventive Principle:
Principle #3Local quality

3Difficulty of detecting and measuring

If histogram analysis and threshold determination are performed to enable selective colorization, then object highlighting is improved, but computational complexity increases

Engineering Contradiction:
Improveobject highlightingVSAvoidprocessing complexity
Core Design Contradiction:
Difficulty of detecting and measuringVSDevice complexity

Solution Approach 1:

The patent performs preliminary actions by first generating a histogram of pixel intensities and determining appropriate thresholds before actually applying the colorization. This preliminary analysis enables efficient selective colorization in subsequent steps, as the threshold criteria are established in advance based on the image's intensity distribution, reducing the need for complex real-time calculations during the colorization process itself.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4214684B1Selective colorization of thermal imaging
Publication Date: 2024.09.04 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP4214684B1 patent drawingFigure 1
  • EP4214684B1 patent drawingFigure 2
  • EP4214684B1 patent drawingFigure 3

AI summary

An example computing system comprises a processor and a storage device holding instructions executable by the processor to receive a thermal image acquired via a thermal imaging system, each pixel of the thermal image comprising an intensity level, and generate a histogram via binning pixels by intensity level. The instructions are further executable to, based at least on the histogram, determine a subset of pixels to colorize, colorize the subset of pixels to produce a selectively colorized image, and output the selectively colorized image.