摘要
蔬菜植株面积大小是评估其生长状况的重要依据。本文采用基于机器视觉的图像处理技术,获取并分析蔬菜的生长状况,其关键在于如何准确地分割图像并计算叶面积大小。为此,通过对采用传统的阈值分割算法、传统的K-means分割算法和L*a*b*空间下的K-means分割算法所输出结果的对比分析,结果表明在L*a*b*空间下进行的K-means分割,在保留蔬菜叶片表面信息的同时不仅有效地分割图像,而且能以彩色图像的形式输出。本文选取了30株绿色蔬菜,计算其在第15天、30天、45天的叶面积大小,通过对比同一蔬菜在不同时间与同一时间及同期内不同蔬菜的叶面积,评估得到这30株蔬菜的生长状况。
Plant area is an important basis for evaluating the growth of vegetables.The growth status was collected and calculated by image processing technique based on machine vision to evaluating,segmenting image and calculating leaves area accurately was crucial.Comparison among the results obtained by the traditional threshold segmentation,the traditional K-means clustering segmentation and K-means clustering segmentation in L＊a＊b＊space.The results showed that the surface details of the leaves could be retained,segmented the image effectively and the image could be outputted in colorful form by K-means clustering segmentation in L＊a＊b＊space.The leaves area were calculated in the 15 th,30th,45 th day with 30 selected green vegetables,the growth status of these 30 vegetables was valued by comparing the leaves area of the same vegetable at different time or the different vegetables at the same time.
出处
《中国农机化学报》
2016年第10期162-165,共4页
Journal of Chinese Agricultural Mechanization
基金
国家自然科学基金(51375084)
中央高校基本科研业务费专项资金资助(15D110316
15D110326
15D110314)
