Three-dimensional Biological Microorganism Identification And Recognition Based On Integral Imaging

Integral imaging has been the most interesting three dimensional informationacquiring and displaying technique for its advantages, and it has great significance inthree dimensional object recognition and three dimensional biological microorganismidentification and recognition. The research work of this paper includes theoreticalanalysis and experimental verification of three dimensional recognition based onintegral imaging, design of the integral imaging microscopy device, three dimensionalcorrelation recognition, distortion-tolerant and classification recognition of biologicalmicroorganism. Detailed investigations are as follows.1. A novel integral imaging microscope device is designed. This system canovercome the shortcomings of current imaging systems and offers a new imagingoptical path; a microlens array that matches with optical path is designed using opticalimaging theory and a3-D information acquisition system under weak illumination isdeveloped. This system can retain and fully use the precise imaging structure of themicroscope, and the magnification increased several times, and meanwhile more subtle3-D information of samples is acquired by using diffuse transmission element to obtainthe scattered light information of the sample. Biological microorganism experimentaldata of element images are acquired by this system, and3-D structure and informationof the sample are analyzed qualitatively and therefore it is efficient for this system toacquire the3-D information of the biological microorganism.2. Three dimensional object correlation recognition of biological microorganism isresearched and realized by use of the integral imaging microscope device and opticalcorrelation recognition theory. Firstly, three dimensional object correlation recognitiontheory of biological microorganism is analyzed using integral imaging and thecombination of optical matched filtering correlator. Secondly, three dimensional objectcorrelation recognition of biological microorganism is realized along with experimentalverification, and the recognition results are analyzed by the correlation output and theeffectiveness is validated by experiment results.3. According to the characteristic of diversity and difference of biologicalmicroorganism, three dimensional object distortion-tolerant recognition is researchedand realized. Firstly, three dimensional object rotation-tolerant recognition based onintegral imaging are analyzed theoretically and verified experimentally. Compared withother methods, by introducing synthetic discriminant function into the field of3-Dobject recognition, the method can greatly reduce the data amount during the recognition process and computational complexity and enhance the recognitionefficiency. Secondly, by introducing the method into the field of three dimensionaldistortion-tolerant recognition of biological microorganism, three dimensionalrotation-tolerant recognition and three dimensional translation-tolerant recognition arerealized.4. According to the characteristic of obvious differences and complex features ofbiological microorganism in the same class, three dimensional classification isresearched and realized by introducing the distortion-tolerant recognition theory into thefield of biological microorganism and combining with inter-class recognitioncharacteristic of synthetic discriminant function. Firstly, three dimensional objectrotation-tolerant classification based on integral imaging and the combination ofsynthetic discriminant function are analyzed theoretically and verified experimentally,and the correctness and efficientness of this method are validated. Secondly, threedimensional classification of two kinds of biological microorganism is researched andrealized with mould samples as reference objects.