基于视觉技术的光纤陀螺闭合光路光纤缺陷检测方法

Detecting methods for FOG fiber defect in closed light path based on vision technique

针对光纤陀螺内部光纤在装配后光路闭合,无有效手段进行光纤缺陷检查,对光纤陀螺长期稳定可靠工作带来潜在危险的问题,提出利用红外视觉检测技术检查光纤缺陷,分析光纤缺陷图像特征,采用最大熵法进行图像分割,提出了结合直方图特征、缺陷区域形状特征、缺陷边界形状特征提取方法,采用改进的神经网络分类方法,用两层网络进行缺陷分类,识别判断不同类别的光纤缺陷。针对光纤缺陷图像的处理结果表明,该方法能够有效地检测光纤缺陷,对不可接受的缺陷能够100%地正确判断。

In view that there was no effective method to detect the fiber defect because the light path of FOG fiber is closed after assembling, an infrared vision detection technical inspection method was put forward to detect the fiber defect. The fiber defect image feature was analyzed, and a maximum entropy method was proposed to segment the image. An extraction method was proposed by applying the histogram feature and the defect region/border shape features. A neural network classification method was modified and used. By using two-layer network to classify the defect, we can recognize and judge different fiber defects. The processing results of fiber defect image show that the proposed method can effectively detect the fiber defects, and all the unacceptable defects can be correctly judged.