谷物作为主要粮食和饲料作物,在农业生产中占有重要地位。提高谷物产量和质量的主要方法是选育出优质品种,而在选育品种过程中,对种子的计数是不可缺少的内容。目前,种子的计数方法通常采用人工计数,或半自动的光电计数器,很显然传统的计数方法存在着效率低、准确性不高等缺点。笔者利用计算机图像识别和处理技术实现考种过程中的计数,为谷物粒群的识别、精选、分级等的进一步研究奠定必要的理论与实践基础,并为育种工作提供一种高效、快捷的谷物计数新技术
　　用Matlab实现这种方法，既能分步对谷物颗粒图像预处理算法进行仿真测试，又可以很直接地得到谷物颗粒个数的结果。实验证明，用Matlab实现的处理结果比较理想，满足识别的应用性。本文介绍用matlab实现了对谷物颗粒图像的灰度化处理、去除噪声处理、灰度图像的二值化、谷物颗粒的数值计算等。并选取较好的处理步骤和算法参数解决谷物颗粒图像预处理的问题。(The grain plays an important role in the production of agriculture as main food supplyment of mankind and animals. The main way to improve grain yield and quality is to breed high-quality varieties.However, in the process of breeding varieties, counting seeds is an essential element.Nowadays, the counting of seeds is usually done manually,or semi-automatic photoelectric counter. It is obvious that the traditional counting method has the disadvantages of low efficiency and low accuracy. The use of computer image recognition and processing technology to achieve the counting test in the process of selection for identification, grain group, lay the theoretical and practical foundation for further research of the necessary classification, and for breeding work provides an efficient, fast new grain counting technique)