Straight Line Detection via Gradient Histogram Features

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for detecting straight lines in images using the Hough transform are prone to inaccuracies when edge strength is weak or noise is present, leading to broken lines and reduced detection accuracy.

Innovation Solution

The method involves acquiring gradient histogram feature sets for each pixel point, determining candidate directions based on these features, and then determining the precise direction and position of the straight line, without using Hough Transform or binarizing the image, thus avoiding errors caused by binarization.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If Hough transform with binarization is used to detect straight lines, then the detection method is simple and fast, but the detection accuracy deteriorates when edge strength is weak or noise is present

Engineering Contradiction:
Improvedetection speedVSAvoiddetection accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent changes the parameter representation from binary values (0 or 1) to gradient magnitude values. Instead of using binarized images where pixels are either black or white, the invention uses gradient magnitude values that continuously represent the strength of edge detection at each pixel location. This parameter change allows the Hough transform to work with nuanced gradient information rather than harsh binary decisions, improving accuracy while maintaining computational efficiency.

Inventive Principle:
Principle #35Parameter changes

2Speed

If binarization is applied to enhance edge detection, then the processing speed is improved, but the straight line detection accuracy deteriorates due to sensitivity to parameters and noise

Engineering Contradiction:
Improveprocessing speedVSAvoiddetection reliability
Core Design Contradiction:
SpeedVSReliability

Solution Approach 1:

The patent introduces gradient magnitude as an intermediary between the raw image data and the binary decision-making process of traditional edge detection. Instead of directly binarizing pixel values, the method first computes gradient magnitudes which serve as a mediator that preserves continuous information about edge strength. This intermediary representation allows for more reliable detection by maintaining subtle variations in edge intensity that would otherwise be lost in binary conversion.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Measurement precision

If gradient histogram feature sets are computed for each pixel point to improve detection accuracy, then the detection precision is improved, but the computational complexity increases

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

Solution Approach 1:

The patent segments the computational process into two distinct phases: a preprocessing phase where gradient histogram feature sets are computed and stored for each pixel point, and a detection phase where these pre-computed features are used to determine candidate directions and final line parameters. This segmentation allows the computationally intensive feature extraction to be performed once, with the results reused during detection, thereby reducing the overall computational complexity while maintaining high detection precision.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP2911093B1Method and device for detecting straight line
Publication Date: 2017.08.23 XIAOMI INC
  • EP2911093B1 patent drawingFigure 1
  • EP2911093B1 patent drawingFigure 2A
  • EP2911093B1 patent drawingFigure 2B~3

AI summary

The present disclosure relates to a method and a device for detecting a straight line. The method includes: acquiring a gradient histogram feature set of each of respective pixel points in an image, wherein the gradient histogram feature set is configured to reflect straight line characteristics of a local area where the pixel point locates (102); determining at least one candidate direction of a straight line to be detected according to the gradient histogram feature sets of the respective pixel points (104); and determining a precise direction and a precise position of the straight line to be detected according to the at least one candidate direction (106). The present disclosure may avoid the influences of errors caused by binarization sufficiently, and increase the accuracy of the detection of the straight line.