The Technology Of Images Processing And Its Application In Maize Germination Tests

With the extensive application of the computer image processing technology in all kinds of aspects, this technology is used more and more in the agriculture domain. However, there are few research reports about the computer image processing technology in the maize germination tests. Using the maize (zea mays) seeds as the germination material with the growing conditions of lighting for 14 hours at 27°C and dark for 10 hours at 15°C, this research builds a series of comprehensive approaches to th e maize germination tests through the image processing of maize seedling and finally builds the model of maize seedling parameters about characters. Furthermore, through the variance analysis, regression analysis and correlation analysis and so on, this research uses the computer image processing technology in the maize growing forecasting and maize seeds storing. Major findings are as follows:1. Using MATLAB language to carry on the pretreatment to the maize seedling image that is obtained, operates the procedure as follows: (1)Using digital camera gather the seedling image while sprouting; (2)Keep the seedling image in the document with JPEG form; (3)Read the image from the computer before the pretreatment of the image processing; (4)Carry on grey level treatment of the colored image;(5)Choose suitable threshold value and change the grey level image into BW image; (6)segment the useless part in two value image; (7)Obtain the extracting features of object in the image, and reject the false seedling; (8)Export features automatically; (9)Carry on the features analysis of the image.2. Draw to the features of object of the maize seedling, it has established the system of new index .For example, Height, Broad, Area, Major Axis Length, Minor Axis Length, Eccentricity, Orientation, Equiv diameter, Extent,3. The model between maize actual seedling height and seedling parameters about characters was built. The relationship between actual seedling height and height seedling of the maize imageis y = 0.974lx-0.8808, r=0.9936, n=223. The absolute error of the result is less than 1.4mm and the relative error is smaller than 3. 6%.The relationship between fresh weight of seedling, actual seedling height and actual seedling broad of the maize image is y = -0.01041+0.002849x,+0.001553x₂,R=0.9770, n=223.4. It was found that the image of maize seedling height, the percentage germination and actual seedling height have correlation. Using height seedling of the maize image, the percentage germination and actual seedling height in the 4^th day could predict the image of maize seedling height, the percentage germination and actual seedling height in the 7th day. The correlation coefficient of height seedling of Nongda 108 maize image is more 0.90. The accuracy rate of the percentage germination of Nongda 108 is 97.37%. The correlation coefficient of height seedling of Nongda 108 maize is more than 0.89. The correlation coefficient of height seedling of Denghai 3 maize image is more 0.74. The accuracy rate of the...