The Research Of The Measuring Method Of Crop Growth Geometric Parameter Based On Image Fusion Technology

The auto identification of glasshouse crops is one of the important research subjects. In this research subject, we study the application of 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 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 means 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 greenhouse 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 segmentation 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 brief 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 optimum(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 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 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 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 optimized segmentation result is obtained.When measuring the growth status parameters of crops, we can use a method depending on virtual video camera to obtain the geometric parameters in view of the plane objects which have complex shapes. In this method, the perspective projection is simplified to homothetic transformation. Therefore the basic properties of homothetic transformation can figure out the geometrical relationship between object in the virtual video camera and the actual object. It simplifies the computing process greatly.