Development Of Near Infrared Component Of Raw Milk Detectiondevice Based On ARM

In recent years, with the continuous improvement of people’s living standards, the demand to dairy product is also increasing, detection for the quality of milk and dairy products has become the focus of the whole society. Raw milk is a kind of complex suspensions which is made of a variety compositions, which contains a large number of nutrients that human bodies need. As the basis of milk and dairy products, the security is causing concern naturally. Therefore, in the process of raw milk acquisition and management, we should strengthen the supervision of the quality, which can ensure the safety of milk and dairy products.The content of protein and fat in raw milk is an important indicator to evaluate its quality. Therefore, detection work to raw milk is mostly focused on the two important indicators of protein and fat at present. The commonly-used detection methods are chemical detection method, ultrasonic detection method and near infrared detection method. Among them, the result of chemical detection method is accurate and recognized by the national, but its experimental period is too long, damage the sample and operation is more complex; Ultrasonic detection method is fast, But most of the testing instrument noise is large and the experiment result accuracy is low; Near infrared detection is a new detection method and is respected in recent year. And this method has many advantages, such as no pollution, rapid detection and real-time etc. Through comparison of several detection methods, this topic combines embedded technology with near-infrared spectroscopy to do the research for the content of the main component of raw milk, its significance lies in the improvement of existing detection method. To design a portable device which is volume portable, speed, on-line analysis and detect the content of the main components in raw milk.This paper considered the composition characteristics and main components of raw milk, researched the Near Infrared Spectroscopy technology’s features and the process of basic data, ultimately determined the overall design of the device. The core of equipment is Samsung S3C6410 A microprocessor and combined with the software program to control the whole equipment. At the same time, selected halogen tungsten lamp as the near infrared light source of this equipment, since spectral coverage which the light source generates is large and the light is compound light which contains a variety of wavelengths, we should use microprocessor to drive the stepping motor, and then make the sub-optical system work to light. Through optical system, we got 900-1700 nm near-infrared light, will this part of the light to the sample, light and sample which passed diffuse transmission effect carried spectral information of the sample, the integrating sphere made persed spectral information together; Used mamatsu Corp’s G9203-256 D to convert the optical signal to electrical signals. Then use ADS1230 to denoise electrical signal and convert them into analog signals. After all works have done, data will be processed by the software program which microprocessor has loaded. Using data of raw milk from the application of equipment to establish calibration models on protein and fat, compared with the calibration model, equipment obtains the predicted value of protein and fat content in raw milk.In last part of the paper, validated the availability and accuracy of the equipment and calibration models, by means of analyzing the evaluation parameters of correction model, we can learn that its parameters are within the scope of the permit and the calibration model has availability. The paper used ten unknown samples of raw milk to spectral detection, compared with the calibration model has been established in the equipment, finally got the predicted value of protein and fat, by contrasting this predicted value with chemical value, we learned that the relative error between the test results of the device and the result of the standard chemical methods is within 10%, and had good linear relation and transitivity. Thus, equipment with availability and high accuracy, can be used in the detection of main composition content in the raw milk, provides a new method for the detection of milk composition of raw material, has wide application prospect and popularization value.