| 基于Kinect相机的穴盘苗生长过程无损监测方法 |
| |
| 引用本文： | 王纪章,顾容榕,孙力,张运.基于Kinect相机的穴盘苗生长过程无损监测方法[J].农业机械学报,2021,52(2):227-235. |
| |
| 作者姓名： | 王纪章 顾容榕 孙力 张运 |
| |
| 作者单位： | 江苏大学现代农业装备与技术教育部重点实验室 |
| |
| 基金项目： | 江苏省重点研发计划项目（BE2018321）、江苏省高等学校自然科学研究重大项目（17KJA416002）、江苏省高校优势学科建设工程项目和江苏省研究生实践创新计划项目（SJCX18-0744） |
| |
| 摘 要： | 为实现工厂化穴盘苗的无损测量,提出一种基于Kinect相机的穴盘苗生长过程无损监测方法。以黄瓜穴盘苗为监测对象,在穴盘苗正上方架设Kinect相机,获取穴盘苗的彩色图像和深度图像,并进行彩色图和深度图之间的像素匹配;通过对彩色图像进行预处理、阈值分割、形态学运算和连通分量统计,获取穴盘发芽率;同时,由图像分割获取的幼苗轮廓和深度值计算得到叶片中心像素点坐标及其对应的深度,以此得到相机到幼苗叶片中心的高度,结合相机到穴盘格的距离和穴盘高度,实现对穴盘苗株高的监测;将深度图像进行直通滤波、条件滤波、边界保持滤波处理,有效去除穴盘苗周围的背景噪声以及波动幅度大的深度数据,获得幼苗叶片中像素点的有效深度,通过在深度图像中对叶片进行重建实现叶面积分析;基于获取的穴盘苗株高和叶面积建立壮苗指数评价模型。利用穴盘苗生长过程监测数据进行实验验证,结果表明,在发芽后5 d内,发芽率误差不大于1.567%;株高和实际株高之间的拟合优度R2为0.875,RMSE为1.395 mm;叶面积平均误差为2.15%;壮苗指数拟合优度R2为0.958。说明本文设计的穴盘苗监测方法可以实现对穴盘苗的发芽率、株高、叶面积和壮苗指数的无损监测,为工厂化穴盘苗生长过程监测提供了有效的解决方案。
|
| 关 键 词： | 穴盘苗 Kinect相机 发芽率 叶面积 株高 无损监测 |
| 收稿时间： | 2020/4/22 0:00:00 |
Non-destructive Monitoring of Plug Seedling Growth Process Based on Kinect Camera |
| |
| WANG Jizhang,GU Rongrong,SUN Li,ZHANG Yun.Non-destructive Monitoring of Plug Seedling Growth Process Based on Kinect Camera[J].Transactions of the Chinese Society of Agricultural Machinery,2021,52(2):227-235. |
| |
| Authors: | WANG Jizhang GU Rongrong SUN Li ZHANG Yun |
| |
| Institution: | Jiangsu University |
| |
| Abstract: | In order to realize the non-destructive and automatic measurement of factory plug seedlings,a non-destructive monitoring method based on Kinect camera was proposed.With the cucumber plug seedlings,the Kinect camera was set up directly above the plug seedlings to obtain the depth image and the color image,and pixels matching was carried between obtained two kind of images.The germination rate was monitored by preprocessing,threshold segmentation,morphological operations and connected component statistics on color image.The center pixel point and the corresponding depth of the leaf were calculated by using the seedling contour obtained by image segmentation,and depth map,so as to obtain the height from the camera to the center of leaf.Lastly,combined with the distance of the camera to the plug tray,the seedling height of plug tray was monitored.The valid depth image was obtained by an algorithm which combined straight-through filtering,conditional filtering,and boundary-preserving filtering to effectively remove background noise around the plug seedlings and depth data with large fluctuations.Then,the leaf area was calculated by blade reconstruction.Based on the results of the height and leaf area of the plug seedling,the estimation of the healthy index was put forward.The method was verified by monitoring the growth process of plug seedlings.Within five days after germination,the monitored germination rate error was no more than 1.567%.The goodness of fit R2between the monitored plant height and the actual plant height was 0.875,and the RMSE was 1.395 mm.The average error of the calculated leaf area was 2.15%,and the R2of healthy index was 0.958.The results showed that the non-destructively monitoring method for plug seedlings can provide an effective solution for the monitoring of the plug seedlings. |
| |
| Keywords: | plug seedlings Kinect camera germination rate leaf area plant height non-destructive monitoring |
| 本文献已被 维普 等数据库收录！ |
| 点击此处可从《农业机械学报》浏览原始摘要信息 |
| 点击此处可从《农业机械学报》下载免费的PDF全文 |
|
