奶牛图像识别系统的研究

Research on the Cow Image Recognition System

对奶牛图像识别系统作了研究,分析了奶牛图像的二值化处理,对光照过强、奶牛背部颜色等影响下的直方图作了分析,并提出了相应改进措施,给出了最佳阈值确定流程图。为进一步完善二值化结果,提出了以黑色点为种子点作区域增长的算法,还提出对二值化图像的识别方法。

In this paper the cow image recognition system has been researched. The binarization process for cow image was analyzed. The histogram under the influence of strong light, the color of cow back was analyzed and the corresponding measures were given. The flow chart of the determination for the best threshold was given. In order to improve the result of binarization process, the algorithm of regional growth with black points taken as seeds was presented. The recognition method for binary image was given. The experimental results showed that the recognition rate with 86% could be achieved with the proposed method.