Algorithm Research On Auto-dimming And Auto-focusing Of Microscope

With the continuous development and improvement of the image processingtechnology and computer technology, computer-aided diagnosis more and moreattention has been paid, has played an integral role in medical research and clinicaldiagnosis, for enhancing the microscope performance, improve diagnostic efficiency,reduce misdiagnosis rate, reduce the labor intensity of the operator has importantpractical significance. Brightness of the light source and focusing accuracy are the twokey factors affecting microscope image quality. Based on its own research anddevelopment Mycobacterium tuberculosis-assisted detection system, for the problems ofthe existing dimming and focusing method, this paper focuses on auto-dimmingalgorithm, the automatic focus theory and methods, noise suppression strategy anddefinition evaluation function on depth study, and provides a good foundation for thesubsequent image segmentation and recognition. The main work and innovations are asfollows:1. This article describes the characteristics of the optical system and control systemdesign structure, explains the hardware and software structure and system workflow,and provides an experimental platform for further research.2. The brightness evaluation based on the empty area of the microscopeauto-dimming algorithm is proposed. By observing a large number of sample images,this article summarizes the image clarity with brightness variation, verifies thedefinition evaluation function dimming performance, and analyzes the feasibility andlimitations. According to the characteristics of serious defocused image, this articleraises the gray average indirect behalf light source brightness values, and obtains themean gray through artificial training samples, then gets optimal brightness throughunilateral extremum search algorithm. The dimming algorithm applied to the ofMycobacterium tuberculosis auxiliary detection system to prove the effectiveness of thealgorithm.3. According to the lens imaging model, this article studies the basic principle ofthe focus, describes several of the PSF estimation model and the basic method of theautomatic focusing. This article analyzes the point source interference by PSF, providesa theoretical basis for eliminating false peaks of the evaluation function. This articledescribes the common noise model and noise spatial filter. The article uses noisesuppression strategies to enhance the processing speed, and effectively improve thefocus function curve.4. This article analyzes the interference of a point source and the edge pixelsampling; puts forward the idea that increasing sampling pixels interval and joined the squared term. According to function curve characteristics, this article puts forward fivefunction performance evaluations, provides objective evaluation criteria to experiment.And sampling pixels interval determined for Mycobacterium image by experiment.Then compared experiments were taken with Tenengrad and Brenner. The results showthat the function keeping the focus speed, increasing the degree of sharpness of thefocusing region, and focusing performance has obvious advantages.