基于神经区域生长瓷砖表面缺陷检测

自动视觉检测是机器视觉在工业方面的一项重要应用。针对目前瓷砖表面缺陷检测仍停留在手工操作水平,劳动强度大,效率低,检测精度远远不能满足实际生产的需要,本文提出一种基于BP神经网络与区域生长法相结合的图像分割技术,并将其应用到瓷砖表面缺陷检测。本算法原理是通过BP神经网络对瓷砖表面进行缺陷检测,将瓷砖主要缺陷分割出来,然后再利用区域生长法对其缺陷部分作进一步分割,使缺陷能被精确、快速地分割出来。通过大量实验说明本算法在实际应用中的精确度达到97%,检测速度得到明显的提高,效果令人满意,具有良好的应用前景。

Defect Inspection of Ceramic Tile Surface Based on BP Neural Network and Region Growing

Automatic visual inspection is an important application of machine vision in industry. Nowadays defect of the ceramic tile surface is still inspected by man, it is a hard work and has low efficiency, accuracy of inspecting can not satisfy the demand of the practical produce. A kind of BP neural network and region growing based on image segment technology is proposed and this technology is applied in the inspection of defect ceramic tile surface. This algorithm is that the defect of the ceramic tile surface is inspected by BP neural network and the region of the ceramic tile defect is segregated by the region growing, the defect is segregated accurately and quickly. The effect of this algorithm is tested in lots of experiments, accuracy is reached by 97% and the velocity is accelerated evidently, the result is satisfied and practical.