| Color measurement is an important criterion as well as an effective means to evaluate products'quality. It is widely used in food industry. Color sorter is a main apparatus for color measurement. But, because the technology of color sorter is not perfect in our country, many corporations have to import foreign machines. China is a large agricultural nation. So, the development of a color sorter can be well suited to China's actual situation and widely applied in the food and products processing industries. It will bring great economic and social benefits.After studying the international advanced technology and design thoughts, a double CPU master-slave color sorting system is designed in this thesis. Firstly, this thesis studies and analyzes the structure components of the color sorting system. Then, this thesis analyzes the factors which influence color sorting system and offers suitable methods to solve the problems. This thesis emphatically studies and designs the subsystem of the controlling system and the subsystem of photo-electrical detecting system. In the controlling subsystem, ARM is adopted as the master controlling CPU, and DSP is adopted as the slave operation CPU to calculate the signals. In this way, a double CPU structure with master-slave coordination mechanism is designed, which can process data and control the system together. Thus, the rate of operation and the ability of data processing are greatly improved. In the subsystem of photo-electrical detecting system, the photo sensor is consisted of silicon photodiodes and red, yellow and blue filters which cover the photodiodes. According to the theory of tri-stimulus values, a color recognition mode is established by measuring the photo-current of the tri-stimulus on the materiel surface. The recognition method based on neutral network is used to recognize and sort rice color, and the method of using neutral network in DSP is designed. The results indicate that this method can greatly improve the sorting precision. |
PDF Full Text Request