Research And Development Of Microbial Sample Recognition System Based On Machine Vision

Microbial detection is the basic operation in the process of food testing,biopharmaceutical and water quality testing.With the increasing attention to food safety issues,the tasks of relevant food testing departments are also increasing,especially during the epidemic period,the detection of food is more stringent.However,it is difficult to identify the colony count of food microorganism because of its small target and various shapes.At present,there is still a gap in the mature equipment for real-time detection of colonies in microbial samples.Therefore,the research on automatic cultivation and colony count analysis of microorganisms has become a research hotspot of relevant departments.In this paper,machine vision as the main research direction,through the python program,image processing algorithm and the related knowledge of microbial culture and counting,the final completion of the colony count,to achieve the purpose of food detection.First,the microorganisms in the food to be inspected are cultured and collected.Because the collected images are color images,the amount of data processed is very large.To solve this problem,this paper uses Python to process the image of microorganisms to reduce the amount of data to be processed.In order to remove the system noise such as water vapor particles in the microbial image,the various noise reduction algorithms are compared Finally,the median filter is selected to smooth the image;in order to recognize the colony contour and eliminate the culture dish edge,the Sobel operator is improved to complete the image edge detection,which makes the colony contour and the culture dish edge contour more obvious.Because the gray levels of the above processed images are more,which is not conducive to counting research,the improved maximum inter class variance method is used to segment the image threshold,and the binary image is obtained,and the amount of image data is reduced,so that the colony target and the background are more clear;in view of the adhesion between the colony and the colony,the watershed processing of adaptive distance transformation is carried out in Python To realize the segmentation of adhesions for subsequent counting processing.Finally,the colony count in the microbial image is completed by the contour detection method based on python.Compared with the standard manual counting,the average error is about 10%,which meets the requirements.Then calculate the colony concentration,and then according to the colony coverage area analysis it is proportional to the concentration.Finally,the hardware design and physical construction of the integrated equipment of microbial culture and detection are completed,and the software part of the equipment is developed on the basis of it.The software part is mainly the interface development and image processing system based on Python language.The equipment designed in this project can not only realize the function of constant temperature cultivation of microorganisms,but also combine with image processing technology to enable the operator to monitor the growth and distribution of microorganisms in the culture device at the computer terminal.Then,the statistics of the colony number and the calculation of the colony concentration in the microbial image are completed,which is convenient for researchers to further study.The purpose is to complete the counting of the bacterial colonies in food microorganisms,and finally display the counting results in the system interface.In conclusion,the combination and improvement of different image processing algorithms,as well as the preparation and operation of software program will become the research focus of this topic;colony adhesion segmentation will be the research difficulty of this topic.