|
Rape as one of the world's four major oil crops, widely grown around the world. There are a variety of colors of rape seed, roughly divided into several black, brown and yellow. The same genetic background, the type of yellow-seeded rapeseed relative to the traditional black seed, brown seed varieties, seed coat thickness, low hull, high oil content, good meal protein content and quality of the oil, cellulose and and low polyphenol content of a series of advantages, yellow seed cultivar selection has become one of the current main research direction of rapeseed breeding. The canola seed color trait relative to other traits of rapeseed, with complicated color gradient diverse, and thus the distinction and classification of canola seed color from yellow to black, there are some difficulties. The conventional method of identifying multi-subjective distinction, identification and classification of yellow seed black seed grading standards is more difficult to determine, the lack of uniform standards of quantity and automation. In this study, the grain color image acquisition information into computer grain color values, combined with near-infrared spectroscopy, and strive to have a qualitative and quantitative study on rape seed grain color traits. As the analysis process is simple, fast, and does not destroy the sample and advantages, so a wide range of applications in the field of agriculture in today's near-infrared spectroscopy (NIRS). Similarly, through the many years of accumulation, NIRS spectroscopy study on rape seed has been relatively mature results. However, these studies are mostly confined to the rape seeds of several quality traits, such as: oil content, protein content, fatty acid content, cellulose content, glucosinolate content and so on rape seed grain color traits related Few reports. Near-infrared spectral data of this study to oilseed rape Engineering Research Center of Southwest University, Chongqing rape seed material, the use of near-infrared spectroscopy, measuring 306 Brassica napus samples and 136 parts of Brassica campestris. Samples transformed the seed grain color image scan acquisition and grain color image to the the grain color values ??(RGB value of the form), the primary color values ??in the rape seed color near infrared spectroscopy mathematical model, through computer processing means combined with near-infrared spectroscopy, and laid the foundation for further study for rape seed grain color traits, and also provides a new reference level of identification with the division of the rape seed grain color. The main results of this study are as follows: in Brassica napus and Brassica napus material grain color, R values ??that best reflect the of rapeseed color traits change with reference to visual observation 0-5 grade samples were divided into six grades, R value corresponding to each level the grain color image of the sample information: 0 less than level 55 to 100,2 100 55,1 to 125,3 125 to 140,4 140 to 150,5 level greater than 150. Construction of R values ??NIRS mathematical model best effect, the internal cross-checking the correlation coefficient (1-VR) is greater than 0.85; an external examination of the correlation coefficient (RSQ) is also greater than 0.83. R-value in the use of the grain color information value built near infrared spectroscopy mathematical model has high stability and accuracy, rape seed grain color to provide a basis for the qualitative and quantitative analysis provide quantitative criteria for canola grain color grading, provide certain reference value for the yellow SEEDED the breeding.
|