Impurity Detection In Liquid Based Image Process

The detection of impurity in liquid is important in the liquid detection. At present, those detection is manual work or via High Performance Liquid Chromatography (HPLC) device. Using manual method is complicated. It wastes much time and manpower. And there are many factors influencing the detection result, such as difference of environment (lightness or hue), manipulator's vision fatigue for working continuously. However, the most HPLC devices are made overseas and very expensive. And the relevant after service is few inland, so repair is difficult.In view of the above mentioned problem, the paper presents some methods. First, get the image of the liquid, then analyse the image using computer graphic processing technology to make out the impurity result. The method is expected to count the impurity thickness efficiently and quickly.The liquid sampling is the first step of manual detection. Detection swatch is taken out from the top, middle and bottom of container. This sampling method can give attention to deposit, suspending or floating dust which is existed possibly. Then detection swatch is put in cuvette or other transparent vessel to examine. Using those experiences for reference, in this paper the sampling tactic is same to manual sampling.Because we need to take image of detected liquid, designation of special detecting device is necessary. The device expected to be designed to meet the desire of imaging valid. And the other hand, the device should reduce the distortedness and transformation. According to the fact, the author shows the device designation in detail and every specification including material, shape, and setting of light and standard of image camera. Especially, the author takes into account the position of light and image camera.While image is acquisited, influence of device and environment is inevitable. So image preprocessing is the chiefly, which includes reducing of noise, illumination and shadow and so on. Those processing can give a clean image for later impurity detection. The paper enumerates the common color space and chooses RGB and HSI. After researching present theory, the author finds arithmetic or method and does experiment to compare.The most important is impurity detection. After studying edge detection in color image and learning present methods to calculate color distance, we analyse the bizarrerie of color space. Then find the way fitting the title to calculate color distance. if the distance between current pixel and background pixel is larger than,the current pixel is marked as impurity.At the end,count the liquid impurity.Complete the software design and describe the definition of mostly class and function.Then, expriment is done.The paper finds the method to resolve liquid impurity detection using computer graphic technology, which is base to improve and optimize.