The Design Of Lensless Folding Reflective Planar Micro-Optical Correlator

Optical Correlators(OCs) have such advantages,there are high-speed, high-precision, parallel data processing and anti-electromagnetic interference, and OCs also have been extensively used in pattern recognition, machine vision, mass information retrieval, information security and artificial intelligence. But the OCs, like typical planar structure and Vander Lugt configuration have many defects, such as too many components, large in size, difficult to integrate, and small in distortion tolerance. According to the research on micro-optics and optical information processing in our laboratory, I use the technology of optical information processing and micro-optics in the design of optical correlator. This design proposes a planar micro-optic correlator which is lensless, and folding reflective. To improve the defects of OCs, I design the typical planar structure of the optical correlator into the planar micro-optical correlator with lensless and zigzag, and use spatial light modulator which combine the digital micro-lenses with target and filter. The planar micro-optical correlator achieved a structure of optical correlator with non-physical lenses, and the correlator can also compress the optical path and volume of the system, increase the integration of devices effectively. The corresponding Optimal Trade-off SDF(OTSDF) filter has also been designed, and achieved distortion invariant pattern recognition within a certain range. The main research work of this paper is as follows:(1) Having researched the basic principles of a typical optical correlator, having deeply discussed the works of the traditional coaxial 4f optical correlator, the typical planar optical correlator and existing lensless optical correlator, and having analyzed the characteristics of different types of optical correlators.(2) The traditional correlators have some common disadvantages, such as, large volume, separate fixed components and lower performance of the recognition. Our laboratory proposes a structure of planar micro-optical correlator with lensless and folding reflective. The model of system and structure parameters of this optical correlator is analyzed and designed, and the implementation of major components in this system is researched. At the same time, digital reflective phase-only micro lenses to meet the structural requirements are designed. This system does not need physical lenses, thus compresses system volume effectively, the volume of this optical correlator system designed is 3V ?63.17 cm.(3) The simulation model of lensless, zigzag and planar micro-optical correlator is established to analyze the effect of recognizing scaling, rotation target. Based on this system, we have designed Optimal Trade-off SDF(OTSDF) matched filter to adapt to different distorted situation.(4) On the experimental platform, identified target and digital ph ase-only micro lens are loaded on SLM1, the matched filter and digital phase-only micro lens are loaded on SLM2. The correct designing is proved. The experimental results show that the system can effectively identify distorted targets within a scaling of 5 2%~150%, rotation angle of-50 °~ +42 °,or a scaling range of 72% ~130% within a rotation angle of-36 °~+38 °.(5) For the matched filter correlator, the experimental results related to the authenticity of the peak are explored, summarized the judgment method by correlation peak intensity, correlation peak shape and rotating the angle of the input image.In conclusion, the planar micro-optical correlator with lensless and folding reflective is a preferred design. This structure has definite guiding signifi cance for components field of the optical pattern recognition and the development of the optical correlator. Optical correlator presented in this paper doesn’t use physical lenses, optical path is folded. Volume is compressed and planar structure is integr ated, and can be assembled and adjusted easily during use. The optical correlator with ability of a range of distortion invariant recognition can better identify the target and has significant application value.