Auto-focus Technology Of Opticai Microscope For Automatic Detection System Of Food Microbiology

Food is the basic condition for human survival, and the security of food is animportant guarantee to human life and health. The generation and development ofthe modern automatic computer vision detection technology provides a new direction for themicrobial detection technology, which is one of the sources of food pollution. Based on theautomatic computer rapid detection system of food microorganism, this paper havedone an in-depth study to the optical microscopy auto-focus technology, the mainresearch issues include the following aspects:The first, this paper takes the color image as the original image. Based on thecolor image RGB color space, this paper puts forward two new kinds of color imagesegmentation method, one is the color image plane segmentation method, the other is the colorimage Gaussian segmentation method. The two segmentation methods can effectively separate thebackground and the entity of the image.The second, based on the study of existing micro-autofocus guidance functionsand the characteristics of the microscope lens, this paper gives out three new focusingguidance functions: the guiding function of Ostu based on the color image planesegmentation, the guidance function based on the color image Gaussian segmentation,and the guidance function based on the image texture features. Combined with thepractical application, we have a test to the existing and those all new focusingguidance functions, and the results show that those new focus guiding functions canquite accurate point out the best focus position. Those functions always have thosefeatures of a better anti-jamming ability, the large range of response distance,and themore obvious extreme point etc. Compared with the existing methods,they all have amore better performance.The last, according to the features of optical microscope focusing process,in thispaper, we present the SPSA search method of the focus point and the parabolic fitfocusing search method. The practical application results show that those twomethods can find the best focus position quick and accurate from the start positon.