基于图像处理的水稻育秧播种质量检测方法的研究

Study on Seeding Quality Detection Ways for Nursing Seedlings of Rice Based on Digital Image Processing

基于数字图像处理技术,研究出一套适用于水稻播种质量检测的算法。通过投影、差分等方法处理采集到的静态图像从而得到处理目标,然后用自行设计的检测算法判断每穴水稻粒数,最终可计算得到水稻育秧播种的空格率、合格率等。详细介绍了该套算法的具体思路及实现方法,同时通过试验验证了软件检测空格率的相对误差最大为0.2%,合格率最大相对误差为-1.5%,论证了该方案的可行性。

Based on the technology of digital image processing, a set of algorithms which is suitable for seeding quality detection of rice nursing seedlings was developed. Through the projection method and subtraction method, static images collected were dealt with in order to get the objective area, and then the number of rice in every seed cave can be determined by designed detection algorithm, from which the leakage rate and pass rate could be obtained ultimately. In this study the specific ideas of this set of algorithm and the methods how to achieve in details were introduced. At the same time, the maximum relative error of leakage rate and pass rate were acquired through repeated verification of experiments, which was 0.2% and 1.5% respectively. It can be seen from the result that this program was feasible.