基于模糊均值聚类和亮度均衡的钢管自适应计数

Steel Pipe Adaptive Counting Based on Fuzzy Means Clustering and Intensity Balancing

针对现有计数方法的缺陷,根据钢管截面区域与背景图像存在较大亮度差异的特点,提出了基于模糊C均值(FCM)聚类和亮度均衡的钢管自适应计数方法。用亮度均衡等方法对钢管图像进行预处理,降低图像中高光和阴影等的不良影响;利用FCM聚类方法自适应分割图像;对二值图像进行连通区域标记,获取区域几何特征;利用统计学方法和FCM聚类方法剔除非钢管截面区域,统计计数。实验表明,新方法不仅计数速度快,而且计数精度高,同时具有对不同环境条件的自适应性。

Aiming at the shortcoming of existing counting methods, a new method based on fuzzy C-means （FCM） clustering and intensity balancing is proposed for steel pipe adaptive counting, according to that there is a big luminance difference between pipe cross-section area and background image. Firstly, the image of the steel pipe is adjusted and balanced for reducing the adverse effect of highlights and shadows in the image; secondly, the image is segmented by using FCM clustering; thirdly, the binary image is labeled with connected components label to obtain geometrical characteristic of region; lastly, FCM clustering is re-used with statistics to remove the interference region of the binary image, and the accurate number of the steel pipe is obtained. Experimental results show that the accuracy and adaptability of counting are raised by the new method.