Study On Machine-mounted System For Acquisition Of Rape Growing Information Based On Spectral And Vision Image

Precision agriculture is a concept of modern agriculture and a field management tool basic variation. Precision agriculture supported by the information technology position, time, and quantity is the development trend of modern agriculture. Quick access to information on crop growth is the key to the implementation of precision agriculture, which is also the current hot topic in the field of precision agriculture.Rape is the most important oil crops in China. The planting area and output ranked first in the world, but still need to import large quantities of edible oil each year. With high level of mechanization, Jiangsu and Zhejiang are the main producing areas of rape where economy has developed, and there are conditions and demands of developing precision agricultural. Therefore, type of rape to carry out machine-mounted nutrients rapid collection of information technology and equipment research has important theoretical and practical value.The purpose of this research is establishing model for quantitative analysis between the spectrum and the main nutrients of rape information. At the same time quickly get rapeseed visual image through machine vision technology. Then establish the relationship between rapeseed growth information and visual image information. Prepare to achieve non-destructive testing of the rapid growth of information of rape.The main contents of this paper are as follows:1) Study the spectral characteristics of rape leaves by using the visible and near infrared spectroscopy in the laboratory. Different calibration methods were developed with whole wavelength region. The results indicated that the prediction precision was optimal by least squares support vector machine (LSSVM) model. The correlation coefficient (r) and root mean squares error of prediction (RMSEP) were 0.8356 and 1.8125.2) Research the rape image processing technology and method, including threshold segmentation, contour extraction and morphological transform. These methods solved the problem of how to accurately draw and calculate the feature parameter of rape. 3) The software of rape quality detection based on visual image was developed. Visual Studio2005 was the development platform, including image geometry transform, morphological transform, rape stopper parameters extraction, rape disease spot parameters extraction and video operation. Achieve the real-time data acquisition, rapeseed image processing, collect information and characteristic parameters extraction, and other functions.4) Design machine-mounted rapeseed growth information detection system, including solar power system and field image acquisition system. Solar power system provides power for the field, includes solar panels, solar Controller and battery. Field image acquisition system includes camera, wireless microwave transmission equipment and image acquisition card. Realize the entire process, get image from the field, then transmission to a computer through wireless and store.5) Research the rape seed image processing technology and method. Calculate the number of rape seed, size and perimeter, extract 5 seed shape parameters, including NCI, circularity, round shape factor, equivalent diameter and shape complexity, and then calculate the average and variance.