Autonomous Driving HDR Imaging for Tunnel Contrast Recognition

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Autonomous driving vehicles face challenges in accurately recognizing geographic features in environments with high brightness contrast, such as tunnels, which can hinder safe predictive driving due to difficulties in understanding the illumination levels inside and outside these areas.

Innovation Solution

An image processing apparatus that generates a synthesized image by adjusting the brightness of objects with significant contrast, using a high dynamic range (HDR) scheme and exposure adjustment based on the difference between the target object and its surroundings, to enhance recognition and reduce unnecessary resource consumption.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If exposure adjustment is performed on all regions of the input image, then the recognition accuracy of target objects is improved, but the computational complexity and processing time increase

Engineering Contradiction:
Improverecognition accuracyVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies exposure adjustment selectively only to regions containing target objects rather than uniformly to the entire image. The image processing apparatus identifies target objects and performs brightness adjustment localized to those specific regions, thereby maintaining recognition accuracy while significantly reducing the computational burden associated with processing the whole image.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent segments the input image into multiple regions based on target object detection. By dividing the image processing task into region-specific operations, the system applies computational resources only where necessary, reducing overall complexity while preserving the recognition accuracy needed for autonomous driving safety.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If exposure adjustment is performed on all frames, then the recognition accuracy of target objects is improved, but the processing time increases

Engineering Contradiction:
Improverecognition accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs exposure adjustment only on frames and regions containing target objects, rather than processing every frame uniformly. This selective approach maintains accurate recognition of tunnel entrances and exits while reducing the total processing time by eliminating redundant computations in frames without relevant objects.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent applies exposure adjustment partially - only to the extent necessary for recognizing target objects, rather than applying it excessively to all image data. This partial action approach ensures sufficient recognition accuracy for safety-critical objects while minimizing processing time delays.

Inventive Principle:
Principle #16Partial or excessive action

3Measurement precision

If brightness adjustment is applied to enhance target object recognition, then the recognition accuracy improves, but the energy consumption increases

Engineering Contradiction:
Improverecognition accuracyVSAvoidenergy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent applies brightness adjustment locally only to regions containing target objects rather than to the entire image. This localized processing maintains the recognition accuracy needed for safe autonomous driving while significantly reducing the energy consumption associated with processing and adjusting the complete image data.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3367303B1Autonomous driving image processing method and apparatus thereof
Publication Date: 2022.09.21 SAMSUNG ELECTRONICS CO LTD
  • EP3367303B1 patent drawingFigure 1
  • EP3367303B1 patent drawingFigure 2~3
  • EP3367303B1 patent drawingFigure 4

AI summary

A processor implemented image processing method includes recognizing a target object in a first frame of an input image; adjusting an exposure of a second frame of the input image based on a brightness of the target object; and generating a synthesized image by synthesizing the first frame and the second frame.