农业工程学报2016，Vol.32Issue(12)：179-186,8.DOI:10.11975/j.issn.1002-6819.2016.12.026
用K-means图像法和主成分分析法监测生菜生长势
Monitoring lettuce growth usingK-means color image segmentation and principal component analysis method
摘要
Abstract
Real-time monitoring of plant growth in greenhouse can provide scientific basis for managing plant production. In order to develop real-time monitoring technology based on machine vision, this paper presents a evaluation method based on image processing and principal component analysis method (PCA) for plant growth. Five independent lettuce plants (S1-S5) and 2 lettuce blocks (G1 and G2) were chose randomly from a greenhouse of a local gardening center. For the single lettuce plant sample, top projected canopy area (TPCA) and plant height (PH) were measured by changing RGB color model to HSI model and by automatic threshold segmentation method. Synchronously, plant height, number of leaf (NOL), length ofX-axis direction of top projected canopy (LX), length ofY-axis direction of top projected canopy (LY), length and width of a certain leaf (LL, WL), which were the six parameters that express a single lettuce growth, were measured manually. The PCA statistical method was used to generate total lettuce growth information (SZS) based on the forementioned six manually measured parameters. Likewise, for the G1 and G2, cover index was calculated based onK-means color image segmentation technology while lettuce plants volume was calculated by the manual measurements. Cover index is defined as TPCA divided by total area of field of view of G1 or G2. Similarly, lettuce plants volume is total volume of the group lettuce plants (G1 or G2). Lettuce growth models were developed for S1-S5 and G1-G2 using regression analysis with higher accuracy (R2>0.80) andP<0.0001, respectively. The results show that there are significant correlation between the total lettuce growth information and image parameters for a single lettuce plant and a group of lettuce plants. These procedures present a good method for assessment of lettuce growth, quantitatively and non-intrusively. The overall results indicate thatK-means color image segmentation and principal component analysis method are feasible for monitoring lettuce plant growth and have potential monitoring many other greenhouse plant growth, on the other hand, the forementioned image segmentation methods and data statistical approach can provide reference for online monitoring of other plant growth.关键词
监测/图像处理/主成分分析/生菜生长势/K-meansKey words
monitoring/image processing/principal component analysis/lettuce growth/K-means分类
信息技术与安全科学引用本文复制引用
李晓斌,王玉顺,付丽红..用K-means图像法和主成分分析法监测生菜生长势[J].农业工程学报,2016,32(12):179-186,8.基金项目
Scientific Research Foundation of Shanxi province (041085) （041085）
Introduce Dr. Scientific Research Foundation of Shanxi Agricultural University (2013YJ26). （2013YJ26）
