IR Image Artifact Detection Using Histogram Regression

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for detecting IR illumination leakage artifacts in IR camera systems are unreliable in uncontrolled environments, particularly due to false classifications from variance analysis and edge-based methods, and deep learning is computationally demanding.

Innovation Solution

A method using frame segmentation, histogram analysis, and regression line fitting to distinguish artifacts from real objects based on smooth gradients, without edge detection, and employing frame subtraction to enhance artifact visibility.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If variance analysis of edge pixels is used to detect artifacts, then the detection process is simple, but false classifications occur because variance loses transition information

Engineering Contradiction:

Improvedetection process complexityVSAvoidartifact detection reliability

Core Design Contradiction:

Device complexityVSReliability

Solution Approach 1:

The patent changes the detection parameter from variance (which loses transition information) to a method that preserves and analyzes the actual pixel intensity transitions. By computing the difference between adjacent pixel values and analyzing the distribution of these differences, the system maintains the transition information needed for reliable artifact detection while keeping the process computationally simple.

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If ESF analysis is used to evaluate edge characteristics, then adequate results are obtained in controlled environments, but the method fails in uncontrolled environments where artifact shapes are unknown or edges cannot be identified

Engineering Contradiction:

Improveedge characteristic evaluation precisionVSAvoidenvironmental adaptability

Core Design Contradiction:

Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent extracts only the essential characteristic needed for artifact detection - the smoothness of intensity transitions - without requiring full ESF analysis or edge identification. By computing pixel differences and analyzing their distribution directly, the method obtains the necessary measurement precision while being adaptable to any environment, including uncontrolled ones where artifact shapes are unknown.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

Instead of trying to identify edges and then analyze their characteristics (the traditional ESF approach), the patent inverts the approach by directly analyzing the intensity transitions themselves. This inversion eliminates the need for edge identification and makes the method universally applicable to all artifact types regardless of shape or environment.

Inventive Principle:

Principle #13The other way round (Inversion)

3Reliability

If deep learning is used for artifact classification, then detection accuracy is improved, but computational demands become too high for practical use

Engineering Contradiction:

Improveartifact classification accuracyVSAvoidcomputational energy consumption

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The patent replaces the expensive, complex deep learning model with a simple, computationally inexpensive statistical method. By using basic operations like pixel difference calculation and distribution analysis, the system achieves reliable artifact classification accuracy with minimal computational energy consumption, making it suitable for resource-constrained embedded systems.

Inventive Principle:

Principle #27Cheap short-living objects (Disposable)

Data Source

PatentUS20250315930A1Method For Detecting Artifacts In Infrared (IR) Illuminated IR Images

Publication Date: 2025.10.09 APTIV TECHNOLOGIES AG

AI summary

The present disclosure relates to a computer implemented method for detecting artifacts in infrared (IR) illuminated IR images. Artifacts are separated from real objects by analyzing results of regression line fits to histograms of pixel values. Artifacts will have smooth transition gradients and therefore will show smaller deviations of the regression line fit to the histogram data.