中文摘要：

针对袋装粮数量智能测算的要求，以实际场景视频图为分析对象，在对典型边缘检测算法实验结果分析的基础上，采用双阈值Canny边缘检测算子和形态学处理的方法提取粮袋特征轮廓——反光带的边界。提出基于众数理论和方差分析相结合的计数算法进行数量的二次判断，提高了计数算法的识别精度和测算效率。实验表明，该方法能准确地提取出粮袋的特征轮廓并进行计数，为粮食数量检测提供了一种有效方法。

英文摘要：

According to the requests of packaged granary grain quantity intelligent reckoning, took the actual scene video as the analysis object, and used the dual threshold Canny operator and the morphology processing method to extract the grain bag characteristic outline-counter-band of light boundary. Then, presented the quantity second-judgment based on the counting algorithm of integration of the mode theory and the variance analysis to improve the counting algorithm recognition precision and efficiency. Experimental results show that the algorithm can accurately extract the characteristic outline and count the number of the characteristic outline, support an effective method for grain quantity detection.