Improvement Design For Fast Monitor Of Grain Quality Based On NIR

In order to carry out the policy of good price for high quality farm produce so that the farm produce in China can be changed from quantity type to quality type; it is in urgent necessity to solve the problem of how to detect the inside quality of farm produce. Therefore, fast farm produce inside quality detection is of high importance. As near-infrared spectroscopy analysis technique has many benefits such as fast, low cost, no touch towards farm produce, a farm produce inside quality detective equipment based on near-infrared spectroscopy analysis technique is developed in this research institution.In this research, the Data collection, optical system, and Charge Coupled Device (CCD) scanning systems parts of existing near-infrared spectroscopy analysis detection equipment has been further developed based on modularized design, in order to future develop and reduce heat and electromagnetism interference maintenance. The main objectives of this research are listed as following:1. Design and develop a Universal Serial Bus (USB) based high speed data collection and transfer module. A USB interface circuit and a multi-channel data collection circuit were designed and developed using the CY7C68013 chip as the core and integrated with A/D converter chip (ADS7825).2. The software development including lower-end computer program, USB driver program, and spectroscopy data analysis program. The lower-end computer program software contains two components—chip firmware program for data transfer and AD data collecting program for data collection. The spectroscopy analysis program carries out spectroscopy data pre-treatment and quantitive analysis. Spectroscopy data analysis program focuses on spectroscopy first order differential coefficient, second order differential coefficient, wavelet transform filter, Principal Component Regression(PCR), select wavelength using Genetic Algorithm(GA)and partial least square, etc.3. Proposed and developed CCD multi-channel compensational detection method. Different from the original double-channel CCD data collection, this research proposed and developed four-channel to two-CCDl difference data collection. This method reduces the noise caused by temperature and other factors.4. Implementation and experiments. Test has been carried out in USB transfer speed. According to multiple tests, it is proved that the USB transfer section can achieve the data collection requirement of the equipment. Implementations have been carried out on using this renovated near-infrared spectroscopy analysis equipment to detect wheat quality. The accuracy and stabilization of this equipment are demonstrated based on the tests on wheat moisture and albumen. Wavelength are selected by using Genetic Algorithrm.