摘要
为快速准确地检测机器播种性能,提出了一种基于图像处理的播种机播种性能检测方法。首先,通过信号触发,相机连续获取传送带上的种子图像,并将采集到的连续图像拼接成图像片段。运用大津法自动获取二值化图像阈值,通过阈值分割方法提取种子图像,并利用中值法对二值化图像中的种子数量进行清点。基于二值化图像的纵向、横向投影信息,得出种子阵列宽度、阵列中心、种子阵列间距、非种子间隔的坐标信息。基于以上测量参数计算出条播种、穴播种和精密播种性能相关参数。在某试验田进行实际应用测试,结果表明,该方法对机器播种性能的检测是可行的、有效的。
In order to detect the performance of machine sowing rapidly and accurately,a method for the sowing performance detection of the planter based on image processing is proposed.Firstly,the images of seeds on the conveyor belts are acquired by the camera through signal trigger.Then the continuous images collected are put together into the image fragments.The threshold values of the image binarization are obtained automatically with the Otsu method.The images of the seeds are extracted with the threshold segmentation method and the number of the seeds in the binary images is counted with the median method.The coordinates of the width,center and spacing of weed arrays and other spacing information are acquired on the basis of transverse and longitudinal projection information of the binary images.Based on the above measured parameters,the relative parameters of row sowing performance,bunch sowing performance and precise sowing performance are calculated.The practical application test in an experimental field shows that this method for the machine sowing performance detection is feasible and effective.
出处
《测控技术》
CSCD
2015年第2期44-47,共4页
Measurement & Control Technology
关键词
图像处理
机器播种
种子图像
播种机
性能检测
image processing
machine sowing
seed image
planter
performance detection
