An all-optical fiber vibration detection system based on optical fiber laser ring cavity to detect high frequency weak vibration signal is presented. A micro-fiber Mach-Zehnder interferometer with multi-channel comb filtering characteristics is formed by tapering polarization-maintaining fiber by fused taper technology. Combined with the fiber laser ring cavity system, MMZI achieves dynamic response to vibrating sound signals in the frequency range from 200Hz to 20000Hz., and the minimum detection pressure of noise limit is 6.44mPa·Hz-1/2. The sound frequency response range of the unconnected fiber laser system is 200Hz to 4 000 Hz and the sensitivity is 29.36mV·Pa-1. After being connected to the laser system, the sound frequency response range is expanded by nearly 5 times and the sensitivity is increased by 6 times, which provides an effective method for broadband sound detection system and micro optical signal detection technology.

Aiming at different types of CCDs with a finite number of pixels, a spectrometer with low wavenumber, high resolution, and wide-spectrum is designed, which is of great significance to improve the optical performance of Raman spectrometer and meet the needs of integration. A low wavenumber high-resolution Raman spectrometer with an excitation wavelength of 532nm, spectral range of 10~2500cm-1, and resolution of 3cm-1 is successfully designed for 2048 pixels CCD based on Czerny Turner M-type optical structure and cylindrical lens to improve the sagittal aberration. The point diagram, RMS (root mean square) radius and modulation transfer function (MTF) curve of the imaging system effectively prove that its optical performance fully meets the design requirements.Then,based on this optical structure,the CCD is replaced by another type of 1024 pixels CCD. By selecting the grating with a stronger splitting power, optimizing the grating angle, and fine-tuning the position parameters of the focusing lens and image plane, the spectral range of the Raman spectrometer is 10~1800cm-1, and the resolution is improved to 1.8cm-1. These two sets of optical paths make full use of the number of the limited pixels of CCD and effectively achieve high resolution and wide spectrum. At the same time, the position of the optical elements of the two sets of optical paths changes little, which effectively ensures the optical-mechanical integration design requirements of a multi-model CCD spectrometer. The work of this article not only ensures the integration requirements of an optical-mechanical system of a spectrometer, but also meets the design requirements of low wavenumber, high resolution, and wide spectrum.

Cassegrain telescopic system has been widely employed in environmental observation such as in astrophysics and navigation. Currently available Cassegrain telescopic system mainly used aspheric mirror systems in order to gain satisfied imaging quality. However, aspheric mirror systems are difficult to fabricate and assemble. In the present study, an easier to be fabricated global Cassegrain telescopic system was designed by using the optical design software Zemax OpticStudio. The structure consists of a primary mirror with a hole, a secondary mirror, and four auxiliary lenses, which the F is 2.6 and the focal length is 760mm. Its wavelength can be operated from 0.9 to 1.7μm and the linear block ratio about 0.45. MTF within the frequency of 35 lp/mm is great than 0.5, and the size (Φ) is small than 350×800mm. According to the above specifications, the structure has a better imaging quality with better aberration correct capability. The proposed structure can provide a new way for designing Cassegrain reflectional telescopic optical systems.

In view of the fact that the attitude measurement of the air bearing table is an important basis for the control of the air bearing table, in order to realize the accurate and fast attitude measurement of the Three-axis air bearing table, a real-time and high-precision attitude measurement method based on monocular vision is proposed. A robust target shape and layout method is adopted to improve the attitude acquisition accuracy. A process of fitting the center of a circle is proposed, and the target can be tracked in real time and robustly through rough acquisition and accurate acquisition methods. Finally, the position and attitude information of the air bearing table is calculated by using the iterative least square attitude estimation algorithm. The experimental results show thatthe real-time frame of the monocular vision measurement algorithm can reach up to 20 frames, and the attitude measurement accuracy is about 0.02 degrees under static conditions, 0.12 degrees under dynamic conditions and can meet the real-time, accuracy and robustness requirements of the air bearing table for attitude measurement.

Spaceborne coherent wind lidar is an effective method to obtain global three-dimensional wind field. The wind field vector inversion method of the coherent wind lidar under the spaceborne platform is theoretically analyzed, and the spaceborne coherent wind lidar simulation system is established based on the theoretical model of the atmospheric stratification time domain signal. The performance of the detection scheme is evaluated by simulation. After preliminary simulation verification, when the orbit height is 270km and the nadir angle is 35°, the wind profile inversion with a resolution of 11.15km along the orbit can be achieved; the horizontal wind speed error is 0.52m/s, and the horizontal wind direction error is 0.19°.

The OCT is used to preliminary explore the effect of moxibustion on epidermal microcirculation at different points on the meridian. Three moxibustion experiments were carried out for each healthy volunteer, and a total of 30 experimental data were obtained. Each group performed 10 measurements on four observation points, and a total of 1200 data were obtained. Based on the imaging characteristics of OCT signals and the optical characteristic parameters in tissue optics, the backscattered signal interval at the acupoint depth of 0.4~1.0mm is fitted to obtain the light attenuation coefficient of the acupoint skin. The results show that compared with the skin light attenuation coefficient in the natural state, the skin light attenuation coefficient of acupuncture points increases by about 40% within 30 minutes after moxibustion, and the backscattered signal intensity of other acupoints on the meridian point where the moxibustion point was located is significantly reduced. This phenomenon is not observed at non-acupoint control points near the meridians and at other control points on the meridians. The results show that the effect of moxibustion on acupoint skin microcirculation can be transmitted to other acupuncture points through the channels of meridians using OCT. This phenomenon has certain clinical significance and further extends the scope of application of traditional OCT.

Metal is an important component in building metamaterials. Therefore, it is very important to determine the surface second-order nonlinear susceptibility of metal in the study of metamaterial’s nonlinearity. A high-sensitive spectrometer was used to measure the second harmonic generation and the acquisition signal was calibrated by the Maker fringe of the quartz crystal whose second-order nonlinear susceptibility is known. Experimentally, a femtosecond laser with a wavelength of 800nm was used as the laser source. The second harmonic signals generated from a 35nm-thick gold film prepared by vacuum evaporation were systematically measured under different polarization conditions at incident angles from 10° to 80°. Then three second-order nonlinear susceptibilities of the gold film were determined. The experimental results are in good agreement with the theoretical calculation.

At present, machine vision is often used in large-scale repetitive industrial production process and some occasions where artificial vision is difficult to meet the requirements, through the machine vision detection method can greatly improve the production efficiency and automation. For some objects that need multi-faceted high-accuracy detection, such as semiconductor thermoelectric cooler (TEC) components, it is often necessary to configure an image acquisition device for each facet and use multiple sets of mechanisms to realize multi-facet detection, which increases the complexity of installation and reduces the reliability of the device. Simultaneous equal optical path confocal imaging of semiconductor TEC components’ two adjacent surfaces based on two-color separation imaging method is proposed, so as to reduce the number of image acquisition devices and the complexity of the device. The device can realize the same optical path confocal imaging of two adjacent surfaces of the TEC component simultaneously, and can be found applications in the automatic inspection fields of machine vision which needs multi-facet inspection, so as to reduce the cost in the actual manufacturing process.

Recent interest in high-precision machining for mechanical devices and semiconductor devices has created a compelling need for miniaturization of displacement measuring devices with high performance. Here a new method for displacement measurement based on self-imaging effect in a double-optical-micrograting structure is proposed. Optical field distribution behind an optical micrograting with period of~6.5μm is analyzed theoretically. And transmission properties of the double grating are investigated. Sinusoidal signals are obtained within a range of 1mm experimentally. Using a subdivision process with a multiple factor of 100, displacement measurement with resolution of 64.5nm is demonstrated. Benefitting from acompact structure and a high stability, the method shows great potential in applications such as miniaturized displacement measuring devices and systems.

Aiming at the accuracy and efficiency of manual and traditional automatic detection of polarizer surface defects, and solving the problem of poor manual design features and generalization capabilities of traditional machine vision, a polarizer surface defect detection method based on improved Faster-RCNN is proposed. First, by comparing the four feature extraction networks, finally select ResNet-101 and introduce the Feature Pyramid Network (FPN) to improve the detection ability of small defects; then use ROI Align instead of the original ROI Pooling to solve the pixel error caused by two rounding operations; Finally, the polarizer surface image is acquired through the acquisition scheme, and three types of defect data sets are established, and the k-means++ clustering algorithm is combined to improve the anchor generation scheme. Experiments show that the improved network has a mAP of 93.5% in the polarizer defect dataset, and the average detection time for a single image to be tested is 0.142s, which can meet the needs of actual detection.

Aiming at the high-precision and universal measurement of the apex radius of curvature of rotationally symmetrical aspheric optical elements with different quadric constants, a laser differential confocal-based aspheric apex radius of curvature measurement method is proposed. This method uses circular pupil technology to make the measured confocal point of the aspheric element close to the center of the apex spherical sphere. Using the differential confocal technology, the cat’s eye point and the confocal point of the measured aspherical component are precisely aimed and positioned respectively, and the actual vertex radius of curvature is measured. Through theoretical simulation analysis, the relationship between the size of the circular aperture and the theoretical deviation of the vertex radius of curvature is obtained, and the actual measured value is compensated. Furthermore, high-precision and universal measurement of the apex radius of curvature of rotationally symmetrical aspheric elements with different quadric constants are realized. Theoretical analysis and experimental results show that the relative accuracy of this method for measuring the apex radius of curvature of rotationally symmetrical aspherical elements is less than 0.01%. It provides a brand-new technical means for the high-precision and universal measurement of the apex radius of curvature of aspheric optical elements.

Mountainous circular structures can provide abundant information on deep geological processes. To deepen the exploration and research of this special geological structure, based on the GF-5 satellite image. The circular structure in the south of Lufeng Dinosaur Valley is classified by land use/cover classification and then the hyperspectral characteristic of the classified land tapes is analyzed. Finally, the hyperspectral characteristics of the same land types inside and outside the circular structure are analyzed. The following conclusions are obtained: In the study area, the unused land accounts for the largest proportion(about 36.14%),which is mainly distributed in the circular form;Forest land accounts for 29.03%,and grassland accounts for 17.80% of the total area; Compared with the spectral library data, the spectral curves of the six land types are consistent with those of the corresponding features, indicating that the object-oriented classification results are effective; The overall trend of hyperspectral curves of the same land use/cover types inside and outside the circular structure is similar, but there are still differences. The maximum reflectance difference of unused land, woodland and grassland inside and outside the circular structure is 6.61%, 4.33% and 4.05%, respectively, but there are bands with only 0.01% reflectance difference. The results provide a theoretical basis and technical support for the establishment of a spaceborne hyperspectral database of land use/cover types of special geological structures in central Yunnan and also provide basic data for further and fine research based on hyperspectral technology.

Because of the existence of high order dispersion, non-linear distortion of optical signals usually appears during the fiber propagation process, it results in the increase of bit error rate of received signals. In view of this, a prediction algorithm of optical signals based on deep neural networks is proposed. This algorithm summarizes the signal propagation model of fiber communication system first, then decides the parameters of input layer and output layer for the deep neural networks. Finally, based on Marine Predators Algorithm the algorithm searches the parameters of hidden layers of neural networks during the training phase. Experimental results show that the deep neural networks can predict the correct wave of received optical signals, and it can provide foundation for dispersion compensation of fibers.

In view of the nonlinear effects of high-speed fiber communication transmission, a nonlinear distortion compensation method for fiber communication based on convolutional neural networks is proposed. The proposed method takes advantage of classical convolutional neural networks to capture the nonlinear features of signals transmission through fiber, and utilizes a regression layer as the final layer to realize nonlinear fitting for optical signals. Besides, quantum particle swarm optimization algorithm is adopted to search the super parameters of deep convolutional neural networks, in order to reduce the difficulty of convolutional neural networks training. Numerical simulation experimental results show that the quantum particle swarm optimization can optimize the super parameters of convolutional neural networks, at the same time, the trained convolutional neural networks can improve the communication quality of fiber transmission.

The infrared atmospheric transmittance is the basis for analyzing the radiating characteristics of infrared target and infrared radiation introspection calculation. But the atmospheric transmittance calculation software as Modtran is hard to be integrated with infrared radiation characteristics measurement system. Based on the empircal formula, a engineering calculation method of infrared atmospheric transmittance suitable for long distance and arbitrary transmission path is presented. The method stratifies the atmosphere based on geographical location, obtains atmospheric model parameters with measured data, achieves a fast-engineering calculation of moderate resolution atmospheric transmittance in the long-range mid-wave infrared. The calculation results show that the average transmittance gotten by this method is close to Modtran and easy to integrate with the real-time system.

In order to solve the problems of halo effect and detail loss in foggy image enhancement by traditional defogging algorithm, a variational Retinex model based on structure and texture awareness is proposed. The initial illumination component is estimated by guided filter; An improved average local deviation filter is used for image structure and texture awareness, and a weight matrix is designed to improve the variational Retinex model to optimize the illumination component; The inversion image and Retinex theory are used to enhance the fog image; Gamma correction is used to adjust the color of the image. Experiments show that this method can effectively remove the fog in the image, avoid the halo effect, and retain more details.

Aiming at the existing methods being short of effectiveness and robustness for high density impulse noise removal, and failing to preserve and restore the edges and detail structures of image, a twice iterative trimmed mean filter with wavelet detection is proposed. This method takes full benefit of characteristic of impulse noise and the highly correlation of wavelet denoising image, identifies the noisy pixel with extreme intensity values and wavelet denoising image, and uses the iterative trimmed mean filter with 3×3 and 5×5 neighborhood successfully for noisy pixel restoration. The experimental results show that in comparison to existing methods proposed very recently, the proposed method achieves higher PSNR and EPI values, and better visual perception, showing superior capability in denoising and detail structure preservation.

At present, many infrared image enhancement methods use image gray features to enhance the image. These methods ignore the non-uniformity of image gray distribution, resulting in the loss of details and unsatisfactory contrast. In order to overcome the above problems, an infrared image enhancement algorithm based on adaptive histogram equalization coupled with Laplace transform is proposed. Firstly, after the input image is evenly segmented, the degree of non-uniformity of image gray distribution is calculated with the help of Gini coefficient of Lorentz curve. Based on the degree of non-uniformity of image gray distribution, the adaptive upper and lower thresholds are constructed to realize adaptive histogram equalization for image contrast enhancement. Then, the smooth filtering method is introduced to remove the noise in the image. Based on the traditional Laplace transform, the eight neighborhood Laplace transform is formed by integrating the diagonal second derivative information of image pixel value, which is used to sharpen the image edge and other details, so as to enhance the image clarity. Finally, the algorithm is used to enhance different infrared images. The experimental results show that the contrast and clarity of the enhanced image are better, and the visual effect is better.