HDR Object Detection From LDR Cameras in High-Contrast Driving

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Autonomous vehicles equipped with Low Dynamic Range (LDR) cameras struggle to distinguish and recognize objects in scenes with high illumination differences, such as tunnels, sunrise, or sunset, which can lead to accidents.

Innovation Solution

A high-performance object detection system that generates High Dynamic Range (HDR) images using LDR cameras by combining images from multiple cameras with different exposure values, predicting disparity values, and applying exposure fusion and tone mapping techniques to create a unified HDR image.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If LDR cameras are used in autonomous vehicles, then cost is reduced, but object recognition capability in high illumination difference scenes deteriorates

Engineering Contradiction:

Improvecamera costVSAvoidobject recognition capability

Core Design Contradiction:

Ease of manufactureVSReliability

Solution Approach 1:

The patent combines multiple LDR camera images with different exposure values to create an HDR image, merging the advantages of different exposure settings to achieve both cost-effectiveness and high dynamic range performance for reliable object recognition

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent introduces an intermediary HDR image generation process that transforms LDR camera inputs into HDR outputs, enabling LDR cameras to achieve HDR-level object recognition capability without directly using expensive HDR cameras

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If HDR cameras are used to improve object recognition in high illumination difference scenes, then object detection accuracy is improved, but system cost increases

Engineering Contradiction:

Improveobject detection accuracyVSAvoidsystem cost

Core Design Contradiction:

ReliabilityVSEase of manufacture

Solution Approach 1:

The patent creates a virtual HDR image by copying and processing information from multiple LDR camera captures, replicating HDR functionality without requiring physical HDR camera hardware, thus maintaining cost-effectiveness while achieving high detection accuracy

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent changes the exposure parameter across multiple camera captures and combines them to generate HDR images, allowing standard LDR cameras to achieve HDR-level object detection accuracy through parameter variation rather than hardware upgrade

Inventive Principle:

Principle #35Parameter changes

3Reliability

If multiple camera images are combined to generate HDR images, then object separation in high illumination difference conditions is improved, but processing complexity increases

Engineering Contradiction:

Improveobject separation capabilityVSAvoidimage processing complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the image processing task into distinct modules (exposure fusion, tone mapping, object detection) to manage complexity while achieving high object separation capability in HDR images generated from multiple LDR inputs

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS20250078491A1High-performance object detection system using HDR images obtained from LDR cameras in autonomous vehicles

Publication Date: 2025.03.06 ORTA DOGU TEKNIK UNIVERSITESI

AI summary

A high-performance object detection system using HDR images obtained from LDR cameras, which allows for the separation and recognition of detected objects in images under high illumination difference conditions (tunnels, sunrise or sunset, etc.) and prevents autonomous vehicles from causing undesired accidents is provided.