基于机器视觉的精播排种器性能检测方法

Performance Detecting Method of a Machine Vision Based on Precision Seedmeter

介绍了采用机器视觉的精播排种器性能检测方法:通过CCD摄像机直接拍摄动态种子流,对获取的样本进行背景去除、二值化、图像平滑、特征量提取、图像标定等处理;构造了基于排种时间序列的型孔种子数频数统计方法,计算出排种器性能指标。试验表明,此方法检测误差小、速度快。

In this paper, the machine vision method was adopted to detect the performance of precision seedmeter. A dynamic seeding flow was shot by a CCD camera and a series of treatments were proposed including the deleting of backdrop, binary treatment, smooth image, the picking up of exigent values, image standardization, and so on. The frequency statistic method of the number of cell seeds was set up based on the serial time. Finally, the performance of the seedmeter was counted. The experiment showed that the method precision and velocity are satisfying.