Unsupervised polygonal structure fitting method

By using an unsupervised polygonal structural component shape fitting method and optimizing the slope and translation parameters using neural networks, the accuracy and efficiency problems of traditional fitting methods are solved. This enables high-precision, real-time monitoring of railway infrastructure, adapts to changes in dense and sparse point sets, and improves the robustness and efficiency of monitoring.

CN121616596BActive Publication Date: 2026-06-09HANGZHOU HUIJING TECH
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
HANGZHOU HUIJING TECH
Filing Date
2026-02-03
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

In railway infrastructure monitoring, existing technologies suffer from several drawbacks. Traditional polygon fitting methods suffer from accuracy loss, low efficiency, and incompatibility with dynamic changes in sparse/dense point sets. Supervised learning methods, on the other hand, require large-scale labeled data and have poor robustness, making it difficult to meet the needs of unattended, all-weather monitoring.

Method used

An unsupervised polygonal structural component shape fitting method is adopted. The slope and translation parameters are optimized through a neural network model to construct an objective function for polygon fitting. The contour point set is processed by combining farthest point sampling and linear interpolation, and the optimal straight line is generated using a multilayer perceptron to achieve accurate fitting of the polygonal structural component.

Benefits of technology

It achieves high-precision, real-time polygon fitting under dense and sparse point sets, adapts to on-site interference, reduces computational load, and improves monitoring efficiency and robustness.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121616596B_ABST
    Figure CN121616596B_ABST
Patent Text Reader

Abstract

The application discloses a kind of based on unsupervised polygon structural member shape fitting method, belong to railway infrastructure intelligent monitoring and computer vision technical field, by obtaining the contour point set of structural member in image, with slope parameter and its corresponding translation parameter describe the straight line where structure piece edge is, based on the distance between contour point and straight line, construct target function, by optimizing slope parameter and translation parameter, minimize target function to train neural network model, generate the slope parameter and translation parameter corresponding to predicted optimal straight line, obtain the contour point set corresponding to the image of structure to be measured, by trained neural network model, generate optimal slope parameter and translation parameter, have generated optimal straight line, by the intersection point between straight line and straight line, obtain the shape of polygon structural member.The application can accurately fit and state detection to the shape of polygon structural member, to realize the real-time monitoring of railway infrastructure accurately, efficiently.
Need to check novelty before this filing date? Find Prior Art