基于圖像融合技術的作物生長幾何參數(shù)測量方法研究.pdf

1、河北農(nóng)業(yè)大學碩士學位論文基于圖像融合技術的作物生長幾何參數(shù)測量方法研究姓名：寇紅娟申請學位級別：碩士專業(yè)：農(nóng)業(yè)電氣化與自動化指導教師：張曙光2009-06-13The research of the measuring method of crop growth geometric parameter based on image fusion technology Author: Kou hongjuan Supervisor: P

2、rof. Zhang shuguang Major: Agricultural Electrification and Automation Abstract The auto identification of glasshouse crops is one of the important research subjects. In this research subject, we study the application of

3、 computer vision and Pattern Recognition Technique in the field of crop growth. It is very important to measure the crop situation Parameters under every kind of circumstance in the greenhouse. Crop growth parameters in

4、each period are obtained via destructive measurements in the traditional way. To address the issue, the thesis puts forward the measuring method of crop growth geometric parameter based on image fusion technology by mean

5、s of computer vision and image fusion. This method is of practical importance in the control of Greenhouse environment and the Simulation of crop growth process. In response to the complicated background in the greenhous

6、e environment, we propose an effective way with which the image of crop is extracted from background image of Greenhouse environment. This method is a fuzzy C-means clustering algorithm which is presented for image segme

7、ntation based on the ant colony algorithm. Ant Colony Algorithm is a kind of algorithm that simulates swarm intelligence. It has a good performance in solving the problems based on Discrete Space. The thesis has made a b

8、rief introduction on the principle and feature of Ant Colony Algorithm. With the flaw of fuzzy C-means clustering (FCM) algorithm that is difficult to determine the number of clusters and is likely to fall into local opt

9、imum(in the research of image segmentation).combining Ant Colony Algorithm with the fuzzy C-means clustering algorithm, we can make use of Ant Colony Algorithm to get cluster center and the number of clusters. Then use

10、 them as fuzzy C-means clustering algorithm's original cluster center and the number of clusters. It has solved the flaw that is likely to fall into local optimum when the fuzzy C-means clustering algorithm searches

11、cluster center at random. Bringing the original cluster center and the number of clusters which is gained from Ant Colony Algorithm into FCM algorithm, we extract the crop object precisely and improve the algorithm's

12、 rapidity of convergent .The improved algorithm fuses the advantages of the two. if it still has disturbed image after cut, we can remove the background totally with the method of Mathematical Morphology Filter, thus an