Infrared Image Processing via Frequency Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Infrared imaging systems face challenges in effectively enhancing dynamic range compression for infrared images, leading to loss of detail information due to limitations in analog and digital video standards and human perception, which existing techniques like histogram equalization and unsharp masking may not adequately address.
Innovation Solution
The system separates infrared images into three components - detail, background, and locally enhanced components - which are scaled and merged to produce an enhanced output image, optimized for display, while retaining radiometric information and minimizing noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If dynamic range compression is applied to convert 12-16 bit infrared data to 8 bit display range, then the output is compatible with display standards, but detail information is lost
Solution Approach 1:
The patent segments the infrared image into multiple frequency components (low-pass, band-pass, high-pass) and processes each component separately through different enhancement techniques before recombining them. This segmentation allows preservation of detail information in high-frequency components while adapting the overall dynamic range for display compatibility.
Solution Approach 2:
The patent applies local histogram equalization and contrast limited adaptive histogram equalization (CLAHE) to specific frequency components and local regions of the image rather than uniform global processing. This local quality approach enhances details in specific areas while maintaining overall dynamic range compression for display compatibility.
2Illumination intensity
If histogram equalization is applied to enhance image contrast, then visibility is improved, but noise is amplified
Solution Approach 1:
The patent applies histogram equalization selectively to specific frequency components (band-pass and high-pass) rather than the entire image spectrum. By segmenting the processing, contrast enhancement is achieved in detail regions while noise in other frequency ranges remains controlled.
Solution Approach 2:
The patent uses contrast limited adaptive histogram equalization (CLAHE) that applies local contrast enhancement with clipping limits to prevent noise amplification. This local quality approach enhances contrast in specific regions while limiting the propagation of noise through the enhancement process.
3Manufacturing precision
If unsharp masking is used to enhance details, then edge definition is improved, but artifacts are introduced
Solution Approach 1:
The patent implements unsharp masking by separately processing high-frequency components through band-pass and high-pass filtering, then combining them with the low-pass component. This segmented approach enhances edges through controlled high-frequency reinforcement while avoiding the artifacts that result from aggressive global unsharp masking.
4Device complexity
If strict linear conversion is used to maintain simplicity, then processing is fast and simple, but severe loss of detail occurs
Solution Approach 1:
The patent maintains processing efficiency through segmented frequency-based processing that can be implemented with standard digital signal processing operations. By dividing the image into frequency components and applying targeted enhancements, the system achieves better detail preservation than linear conversion while keeping computational complexity manageable through modular processing stages.
Data Source
AI summary
Systems and methods disclosed herein, in accordance with one or more embodiments, provide for processing infrared images. In one embodiment, a system includes an infrared sensor adapted to capture infrared images and a processing component adapted to process the captured infrared images by extracting a high pass part from the captured infrared images, extracting a mid-spatial frequency part from the captured infrared images, extracting a low pass part from the captured infrared images, separately scaling each of the parts, and merging the scaled parts to generate enhanced output images. The system may include a display component adapted to display the enhanced output images.