Luminance Correction for Background Difference Object Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing systems struggle to detect objects in a background difference type when the exposure conditions of a camera change, such as in cases where the illumination varies or the camera moves.

Innovation Solution

The proposed image processing system includes a luminance correction unit that adjusts the luminance values of images based on exposure conditions, allowing for accurate object detection in background difference type even when exposure conditions change.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If background difference processing is used for object detection, then processing load is small and extraction stability is excellent, but when illumination condition changes, the processing cannot be performed accurately

Engineering Contradiction:
Improveprocessing efficiencyVSAvoiddetection accuracy under illumination change
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent applies parameter changes by introducing exposure condition parameters (exposure time, gain) to characterize and correct illumination variations. By modeling the illumination change as a parameter transformation and applying inverse correction, the system maintains background difference processing accuracy under varying illumination conditions while preserving its computational efficiency.

Inventive Principle:
Principle #35Parameter changes

2Reliability

If template matching processing is used to compensate for illumination variation, then detection can be performed under illumination change, but processing load increases

Engineering Contradiction:
Improvedetection accuracy under illumination changeVSAvoidprocessing load
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent extracts only the essential illumination variation component from the complex template matching process. By isolating the exposure condition parameters (exposure time and gain) as the primary sources of illumination change and correcting only for these factors, the system achieves illumination compensation with significantly reduced computational load compared to full template matching.

Inventive Principle:
Principle #2Taking out (Extraction)

3Reliability

If illumination compensation is performed in background difference processing, then detection accuracy improves, but regions other than moving object are incorrectly extracted

Engineering Contradiction:
Improvedetection accuracyVSAvoidbackground difference extraction precision
Core Design Contradiction:
ReliabilityVSMeasurement precision

Solution Approach 1:

The patent introduces exposure condition parameters as an intermediary between the input image and the background difference calculation. By using these parameters to predict and correct illumination variations before difference computation, the system achieves accurate compensation without introducing false differences in static regions, thus maintaining both detection accuracy and extraction precision.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3916675B1Image processing system and camera system
Publication Date: 2025.05.14 ASTEMO LTD
  • EP3916675B1 patent drawingFigure 1
  • EP3916675B1 patent drawingFigure 2~3
  • EP3916675B1 patent drawingFigure 4

AI summary

Provided are an image processing system and a camera system capable of appropriately detecting an object in a background difference type even when an exposure condition of a camera changes. The camera system includes an image processing system 100. The image processing system 100 includes an original image acquisition unit 101 that acquires a first original image and a second original image, a luminance correction unit 104 that acquires a luminance-corrected image by correcting a luminance value of the second original image such that a difference from a luminance value of the first original image decreases in accordance with an exposure condition related to each of the original images, and a background difference type object detection unit 105 that detects an object by acquiring a background difference based on the luminance-corrected image.