Target Edge Separation for Overlap-Robust Image Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current camera systems in autonomous vehicles struggle to accurately distinguish targets from unwanted objects due to edge connections or overlaps, especially when targets and objects have similar textures and colors, leading to inaccurate recognition.

Innovation Solution

An image processing method and apparatus that separates edge points into a primary and secondary class based on height and deviation from a linear regression line, allowing for more accurate target detection by forming a final primary set for target creation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional image processing methods are used for target detection in complex scenes, then the detection process becomes computationally intensive and time-consuming, but the detection accuracy and robustness deteriorate due to false positives and negatives

Engineering Contradiction:
Improvetarget detection accuracyVSAvoiddetection time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent divides the image processing task into multiple stages: obtaining initial detection results, generating candidate regions, performing verification, and filtering false positives. This segmentation allows the system to process images efficiently by focusing computational resources on promising regions rather than analyzing every pixel uniformly, thus reducing detection time while maintaining accuracy.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary actions by pre-processing images to enhance target characteristics before main detection, and by generating candidate regions based on initial detection results. This preliminary filtering and enhancement reduces the computational burden of subsequent verification steps, enabling faster detection without sacrificing accuracy.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If multiple verification steps are added to improve detection accuracy, then the detection robustness improves, but the system complexity and computational load increase

Engineering Contradiction:
Improvedetection robustnessVSAvoidprocessing system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies different processing qualities to different regions of the image. Candidate regions identified as potential targets receive intensive verification processing, while other regions receive minimal or no processing. This local quality approach ensures high detection robustness for actual targets while avoiding unnecessary computational complexity in regions where targets are unlikely to be present.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent applies verification processing selectively to only those regions that are candidates for containing targets, rather than applying uniform verification to the entire image. This partial action approach maintains high detection robustness for potential targets while significantly reducing overall system complexity and computational load compared to exhaustive verification methods.

Inventive Principle:
Principle #16Partial or excessive action

3Measurement precision

If traditional object recognition algorithms are used, then the method is simple to implement, but the detection precision and anti-interference capability are insufficient for complex backgrounds

Engineering Contradiction:
Improvedetection precisionVSAvoidalgorithm complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the detection process into distinct phases: initial detection, candidate region generation, verification, and false positive filtering. Each phase uses appropriately complex algorithms matched to its specific task, allowing the system to achieve high detection precision through coordinated simple and complex operations rather than requiring all components to be highly complex.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces intermediary steps between simple initial detection and final target confirmation, including candidate region generation and verification processes. These intermediary steps act as mediators that refine the detection results, improving precision by filtering out false positives while managing algorithmic complexity through a structured multi-stage approach rather than requiring a single complex algorithm.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3979196B1Image processing method and apparatus for target detection
Publication Date: 2026.05.06 GREAT WALL MOTOR CO LTD
  • EP3979196B1 patent drawingFigure 1~2
  • EP3979196B1 patent drawingFigure 3~4(b)
  • EP3979196B1 patent drawingFigure 5~6(a)

AI summary

The present invention relates to the technical field of intelligent transportation and image processing, and provides an image processing method and apparatus for target detection. The method comprises: performing image preprocessing to obtain an edge point set for a target; in a same edge point set, if there is an edge point connection or overlap between the target and a peripheral edge line of an object, selecting edge points from the edge point set to form a first class and a second class, wherein the first class comprises edge points formed as a primary set and having a height less than the height of the target; performing linear regression processing on the edge points in the primary set to obtain a linear regression line; selecting an edge point from the second class one by one to calculate a deviation thereof with respect to the linear regression line; if the deviation is less than a preset standard deviation, adding the selected edge point to the primary set to form a final primary set; and creating a target on the basis of the edge points in the final primary set. In the present invention, the edge set connected or overlapped in an image is successfully separated.