微型机与应用
微型機與應用
미형궤여응용
MICROCOMPUTER & ITS APPLICATIONS
2012年
23期
45-47,54
,共4页
疵点检测%LBP%SOM%图像金字塔
疵點檢測%LBP%SOM%圖像金字塔
자점검측%LBP%SOM%도상금자탑
defects detection%LBP%SOM%image pyramid
提出了一种联合LBP与SOM的多分辨率织物疵点检测方法。首先应用图像金字塔将织物图像进行降级分解，得到不同分辨率的图像，再对每一级的图像分别应用局部二进制（LBP）算子进行处理，提取特征后送入事先训练好的自组织映射神经网络（SOM）进行识别，最后再对已识别的多级图像进行融合，计算连通区域的周长和面积去除孤立点后得出最终的检测结果。实验结果表明．该方法检测速度快、检测效果好，适用于不同疵点类型的各种检测。
提齣瞭一種聯閤LBP與SOM的多分辨率織物疵點檢測方法。首先應用圖像金字塔將織物圖像進行降級分解，得到不同分辨率的圖像，再對每一級的圖像分彆應用跼部二進製（LBP）算子進行處理，提取特徵後送入事先訓練好的自組織映射神經網絡（SOM）進行識彆，最後再對已識彆的多級圖像進行融閤，計算連通區域的週長和麵積去除孤立點後得齣最終的檢測結果。實驗結果錶明．該方法檢測速度快、檢測效果好，適用于不同疵點類型的各種檢測。
제출료일충연합LBP여SOM적다분변솔직물자점검측방법。수선응용도상금자탑장직물도상진행강급분해，득도불동분변솔적도상，재대매일급적도상분별응용국부이진제（LBP）산자진행처리，제취특정후송입사선훈련호적자조직영사신경망락（SOM）진행식별，최후재대이식별적다급도상진행융합，계산련통구역적주장화면적거제고립점후득출최종적검측결과。실험결과표명．해방법검측속도쾌、검측효과호，괄용우불동자점류형적각충검측。
In this paper, a muhi-resolution fabric defects detection based on LBP and SOM is introduced. Firstly, using image pyramid to downgrade decomposition for the fabric image, and different resolution images was get. Then, LBP is applied for every different resolution images, which are sent to prior trained self-organizing map neural network (SOM) to identify after feature extraction. Finally, it fuses the identified multi-level image, calculates the connected region of perimeter and area of removal of isolated points. The experimental results show that the algorithm has high detection speed and effect, and can be applied to a variety of different types of defect detection.
