摘要
不同类型地板的疵点检测系统需要不同的检测算法,这样不仅会提高生产成本,还会造成产品检测流水线的复杂程度;为此提出了一种对两种不同纹理表面地板疵点的检测方法。首先采集若干幅地板图像,利用光度立体测量技术进行处理,再通过SVM分类器对图像分类。对表面光滑的LC11软木地板疵点进行一般的阈值分割算法处理即可;对于表面粗糙的荔枝纹pvc运动地板预先进行平滑滤波、FFT处理,再通过阈值分割等算法处理。试验结果表明,这种疵点检测方法可以较好地满足对两类地板疵点的检测。
The defect detection of different types of floor systems require different detection algorithm,which will not only improve the production cost,also can cause the complexity of the products pipeline,for this,a kind of surface of two different texture floor defect detection method was put forward.First image acquisition two floor,using the photometric stereo technology for processing,again by the SVM classifier for image classification.On the floor of the smooth surface defect general threshold segmentation algorithms which can;for the floor of the rough surface FFT processing first,then by threshold segmentation algorithm processing,etc.Test results show that this kind of defect detection method proposed can well satisfy the two types of floor defect detection.
出处
《科学技术与工程》
北大核心
2017年第30期263-267,共5页
Science Technology and Engineering
