基于两阶段增强密度聚类的铝合金表面划痕视觉检测方法

打开文本图片集

中图分类号：TP391.41 文献标志码：A文章编号：1006-0316（2025）11-0059-07

doi：10.3969/j.issn.1006-0316.2025.11.009

A Visual Scratch Detection Method for Aluminum Alloy Surfaces Based on Two-Stage Enhanced Density Clustering

SUN Qi， HUANG Jianhong， ZHANG Bo，XU Bin （School of Mechanical Engineering， Sichuan University， Chengdu 61oo65， China）

Abstract ：Asavital metallic material，aluminumallyis widelyused inmodern industries.However，due toits complex preparation processes and intricate manufacturing procedures，surface defects such as scratches are inevitable，which severely affect its production quality and operational performance.This paper proposes a visual inspection system for aluminum plate surface scratches based on a two-stage enhanced density clustering method.The system addresss issues of uneven illumination and texture noise in images by constructing an optimized filtering approach. Furthermore，an enhanced clustering objective function and measurement strategy are developed to enable accurate scratch identification and geometric measurement.Experimental results based onthedetection of different types of surface scratches on aluminum aloys demonstrate that the proposed method can accurately identify and detect various scratch morphologies. It achieves arecognition rate of 100% （20 andarelative error in length measurement of less than 0.97% ，indicatingbroad application prospectsin industrial inspection.

Key words ∵ aluminum alloy i scratch detection ;image filtering i image clustering

铝合金作为一种重要的金属材料，广泛应用于电子信息、航空航天、机械制造等领域。（剩余6701字）