Hemispherical resonator gyroscope(HRG) is a kind of angle rate sensor with high precision, high reliability, and small size weight and power(SWaP). HRG innovative approach is a real game changer in high end navigation, which may change the territory of inertial navigation industry. In this paper, the current advances of HRG are reviewed, the research gap between home and abroad is analyzed. The key technology of HRG is summarized, and the technologies of the precision processing, chemical etching, frequency balancing, film deposition, assembling, electronics and control, error compensation, etc. are emphasized. In the end, the developing trends of HRG technology and market are discussed.

A phase unwrapping method based on feature extraction of binary grating is proposed in this paper. In traditional phase measuring profilometry based on sinusoidal grating projection, the phase is wrapped between (-π, π］due to the arctangent operation, and it needs to be unwrapped using relative phase unwrapping algorithms such as diamond-type phase unwrapping algorithm in which the accumulation of phase errors is easy to occur or even lead to wrong phase unwrapping; or with the assistance of some additional structured light patterns such as Gray codes, the fringe order may be determined to achieve phase unwrapping without the phase error accumulation. By the proposed method, three-dimensional measurement is achieved by just projecting a set of phase-shifted binary gratings, which can not only avoid the Gamma non-linearity influence of the projector, but also increase the refresh frequency of the projector by more than ten times. At the same time, by using the image characteristics of the binary grating itself, the fringe order can be extracted without increasing any auxiliary structured light pattern. So that the unwrapped phase can be achieved by combining the phase of extracted fringe order with the wrapped phase without the phase error accumulation.The experimental results verify the feasibility and effectiveness of the proposed method, and the solution result is in good agreement with the solution result of the diamond algorithm without cumulative errors.

3D facial expression capture is widely used in film and television, game, medicine, social intercourse and other fields. In order to realize high-precision acquisition of 3D facial expression information in unmarked free movement, a fringe projection 3D face measurement method based on key point detection and matching is proposed. Firstly, the quasi-static 3D information of the object to be measured is obtained as a reference model by using three-frequency heterodyne method, and the initial key points is obtained from the modulation image by using the face key point detection algorithm based on deep learning.The 3D coordinates corresponding to these key points are then obtained. During the dynamic measurement, only two or three fringes with phase shifts are projected. Modified Fourier transform profilometry or phase measurement profilometry is used to obtain the wrapped phase and modulation of deformed fringe. Face key points are detected in the modulation image and matched with the initial key points to obtain the coordinate relationship between the key points and the reference model. The motion parameters of the face relative to the reference model are calculated and the approximate phase is further calculated using the reference model. Finally, the approximate phase is used to unwrapthe phase and reconstruct the 3D shape to obtain the 3D facial expression at that time. A 3D facial expression capture system is built by using DLP projector and high-speed camera. 3D data acquisition with 1.3 million pixels at 100 fps is realized. The experimental results show that the system can accurately capture the unmarked 3D expressions with significant changes within ±45 degrees of deflection, ±30 degrees of pitch and ±45 degrees of roll.

A 2+1 phase-shifting algorithm with temporal phase unwrapping is proposed in this paper. Traditional 2+1 phase-shifting algorithm plays a good performance in terms of suppressing errors introduced by motion. In order to simultaneously realize temporal phase unwrapping, only one additional auxiliary sinusoidal grating with lower frequency is needed, then the geometric constraints are used to unwrap phase of the lower frequency deformed pattern by combining Hilbert transform to guide unwrapping the phase of 2+1 algorithm. The experimental results show that the proposed algorithm can successfully recover phase information of objects with large-scale depth and intricate surfaces with higher measuring accuracy and higher measuring speed.

Polarization information and orbital angular momentum information are important carriers of optical signals of optical information, as well as important parameters of optical information processing. In this paper, an Archimedes double helix structure is designed on a quartz substrate, and the finite element method is used to calculate the patterns of transmitted light intensity distribution generated by the coupling of the incident light and the metasurface, when the incident light has different orbital angular momentum and polarization information and irradiates the above-mentioned metasurface. According to this pattern, the polarization information of the incident light can be detected. In order to detect the orbital angular momentum of the incident light, a new calculation method is also proposed, which calculates the offset of the intensity center of the transmitted field generated by the coupling of the incident light and the metasurface and uses it to characterize the orbital angular momentum of the incident light. Finally, the relationship between the orbital angular momentum of incident light and the offset of the intensity center is fitted successfully, which realizes the detection of the orbital angular momentum of incident light. This research provides a new idea for detecting the information carried by incident light.

Aiming at the great difficulties in collecting fabric data-sets and defecting fabric detection, an algorithm of fabric defect detection using deep learning combined with traditional methods is proposed in this paper. Firstly, an autocoding network based feature pyramid structure is proposed, which only needs normal samples for learning. Secondly, in the detection phase, multi-model fusion at the same scale is proposed to reduce the false alarm rate and remove the interference of texture noise. The experimental results show that the learning method proposed in this paper has a detection rate of over 98% for linear defects and over 84% for planar defects. It has more application value in practice.

In order to study the scattering characteristics of weak defects on the surface of precision optical elements, a scattering theoretical model of weak defects on the optical surface is established based on vector scattering theory and bidirectional reflection distribution function. The variation of bidirectional reflection distribution function with scattering angle is simulated, and the effects of incident angle, incident light wavelength and defect size on the scattered light characteristics of defects are analyzed. Based on the analysis of simulation data, the influence of light source parameters on scattering characteristics is simulated and analyzed, which provides a theoretical reference for the selection of system parameters for precision surface defect detection using dark field imaging method. The range of incidence angle for defect detection is 30° ~ 50°. In the visible light range, the wavelength range of 380~500 nm is more conducive to defect detection. In addition, by studying the variation law of scattering field when the defect size changes, it provides a reference basis for distinguishing the shape and size of defects.

Wheel set is an important part of the train, whether the key parameters of wheel set tread exceed the limit is related to vehicle driving safety. Therefore, it is necessary to regularly detect the wear condition of the train wheel set. A method for measuring wheel set tread profile is proposed, in which the design of measurement system and improved gray center algorithm are included. On this basis, a simulation wheel set tread detection experimental system is established to test the result. It is shown by the experiments that the mean square error (MSE) of the 1~10 mm block gauge is within ±0.1 mm, and MSE of the 1~5 mm is less than 0.05 mm, which can meet the high precision requirements of industrial measurement.

Whether the surface of the solar panel is intact or not plays a decisive role in its power generation efficiency. The traditional manual detection method, infrared penetration detection method and machine vision detection method all have their own shortcomings. Due to the difficulty of obtaining outdoor solar panel images, small number of samples, high sample similarity and other problems, most of the deep learning algorithms can not well complete the outdoor solar panel defect detection task. In view of the particularity of this task, a defect detection method based on the improved DenseNet network model is proposed. The DenseNet basic network model is selected, L2 regularization is added to the model, the Batch Normalization layer is adjusted to solve the over-fitting problem, and the activation function ReLU function with the SELU function is replaced, which can better alleviate the problem of gradient disappearance and strengthen the robustness of the network. In the final experiment, the accuracy of the training set is 93%, and the accuracy of the test set is 87%. It can effectively detect and distinguish different degrees of damage to the battery board.

Object detection on the ground battlefield is the basis of precise strike, playing a vital role in modern unmanned warfare. The traditional image algorithm is restricted by lighting, weather and other conditions, which can be solved by 3D detection algorithm using lidar. For unmanned vehicles’ detection task on land battlefield, a 3D detection algorithm based on convolutional neural network is proposed in this paper. By optimizing the feature fusion module of VoxelNet, a group of end-to-end efficient networks are designed, and a non-maximum suppression strategy based on distance is improved. Experiments show that on the self-built dataset, original VoxelNet’s AP of vehicle target is 78.53%, while our network performance is 84.11%, which has great value for 3D detection task in the feature military field.

The number and proportion of five kinds of leukocyte in human peripheral blood reflect the health state of human body. Manual examination of leukocyte consumes a lot of manpower for medical?workers. How to use intelligent method to classify leukocytes quickly and accurately is an urgent problem to be solved. The accuracy of leukocyte segmentation is the key to correct classification. In this paper, an improved iterative threshold image segmentation algorithm is proposed, and the minimum distance method for restoring mitotic lines is improved based on mathematical and digital simulation analysis. The accuracy and efficiency of leukocyte segmentation are improved, and the problems of platelet adhesion and unclear leukocyte boundary are solved by these methods. The leukocytes are separated from the complex blood environment, the minimum distance of the lobulated nucleus leukocytes is determined and connected, and then located to each leukocyte to make a data set, finally, it is classified by CNN. After testing, the accuracy of leukocyte segmentation is more than 96%. The experimental results show that the proposed segmentation method is accurate, efficient and practical.

Compared with the traditional non independent steering chassis, the chassis with four-wheel independent drive steering has higher response speed and flexibility, which is more in line with the development direction of unmanned vehicles in the future. Aiming on controlling four-wheel independent drive unmanned vehicle, kinematic analysis is made in three steering modes with steering angles considered as the control inputs. With the system state space model developed and the cost function and constraint conditions of the system designed, the steering controller based on MPC is established. Simulation experiment using Simulink & Carsim shows that the controller can track the given trajectory quickly and accurately for different control schemes at different vehicle speeds, and has good adaptability and robustness while giving full play to the mobility of four-wheel drive independent steering. It provides fundamental base for the further research.

In the process of laser communication, laser guidance, rendezvous and docking in order to achieve the acquisition and tracking of long-distance targets, a laser tracking method based on single-element detector is proposed in this paper. A single-element detector is used to measure the distance information, and a fast mirror is used to scan the field of view in spiral line. With the angle information of fast mirror and distance information of detector, a three-dimensional image is generated. Through signal processing, the center of target in 3D image can be obtained and the control module drives the fast mirror to start next scan according to the difference between target center and FOV center, so that the target is always in the scanning FOV and target tracking is realized. The advantage of this method is that the single-element detector can be introduced into tracking task, and the array detector is no longer needed to locate the center of target, and therefore the echo energy density is much higher and the tracking range can be much longer, especailly with the incorporation of single photon avanche detector. With this method, the simulation experiment has been carried out and the target at 3.75 m is captured and tracked. The experimental results show that the simulation and experimental results are consistent with each other, and the acquisition probability of target with diagonal velocity of 9.07 mrad/s is 72.5%.

In order to solve the problem of detecting and tracking small infrared targets under the complex sea-sky background, a detection and tracking algorithm for the small infrared targets under the complex sea-sky background is proposed in this paper. In the detection stage, in order to suppress all kinds of clutter in different regions, the algorithm uses different classifiers to distinguish the clutter and small targets in different regions. In the tracking phase, in order to further eliminate isolated noise and clutter interference, the Gaussian mixture probability hypothesis density filter is adopted to track the target. Detecting and tracking experiments on the simulation video show that the proposed algorithm performances better than the previous algorithms, that improves the correct tracking rate about 10%, and the tracking precision about 50% on average. The algorithm has good engineering feasibility.

In the field of shooting range testing, the trigger is mainly used to detect the physical phenomena generated when the projectile is fired, and output the trigger signal at the zero time of the projectile firing to start different subsequent ballistic testing equipment to achieve synchronization of multiple parameters. Existing triggers are easily affected by the environment, and are difficult to handle due to their large size, and have relatively large usage limitations. In order to solve the above limitations, the infrared radiation of the flying projectile in the terminal ballistic section is studied in this paper. A suitable mid-wave infrared detector is selected. A mid-wave infrared trigger is designed. The mid-wave infrared signal is gathered through an optical structure, and is focused on the PbSe detector. Corresponding signal processing circuit is designed on the photosensitive surface to realize accurate detection of projectiles. Tests have verified that the trigger can effectively detect projectiles. It has high detection sensitivity, and has good environmental adaptability and reliability.

Aiming at the fact that the existing infrared polarization image pseudo color technology can only display the collected polarization image scalar and non dynamically, the infrared pseudo color signal processing based on P and S polarization image is studied, and the P and S polarization image information is mapped to RGB color space after processing. The hardware circuit of pseudo color processing system with TMS320DM642 chip as the core processor is designed, and the software design, debugging and analysis of pseudo color signal processing system are completed. By collecting the infrared polarization characteristics reflected from the surface of the measured object, the infrared polarization image processing system obtains more useful information and ensures the processing speed of the infrared polarization image. It is of great significance for infrared target recognition and detection.

The existence of the deviation error of the lens center in the transmissive optical system affects the coaxiality of the optical system, resulting in eccentric aberration and affecting the imaging quality. The application of aspheric lens can reduce the aberration of the optical system, and it is more and more widely used. In order to study the difference of the decentering aberration of the system caused by the decentering error of the aspheric lens and the spherical lens, a method for analyzing the central eccentric error of optical system by optimal optical axis fitting is proposed.?Combined with Seidel polynomials, the misalignment parameters are added to the system by Zemax software, and the difference of the influence of the same center deviation error on the aberration of spherical and aspherical systems is derived. Experimental results show that the aberration of the optical system caused by the deviation of the center of the aspheric lens is relatively large. Compared with spherical lenses, in the processing and manufacturing process of aspheric lenses, the center deviation error must be more strictly controlled.

In order to enable the car to be recognized by other vehicles when driving at night and avoid traffic accidents, a new type of rear position light optical system with a pattern on the condenser is designed. Based on the theory of non-imaging optics, the method of illuminance optimization is adopted and the coordinate value of each point of the free-form surface mirror is calculated by the formula. According to the light distribution requirements of the rear position lamp according to the laws and regulations, the reflecting surface is reasonably divided to meet the light distribution requirements. The light is traced by the Monte Carlo method. The experimental simulation results show that the light distribution effect meets the illuminance requirements of the national regulations GB5920-2008 for each test point of the rear position lamp. The new structure designed makes the light type more uniform and avoids bright occurrence of spots.

The optical system of single infrared spectrum receives less target information when the UAV reconnoitres the ground, and it is also susceptible to the influence of the environment or even unable to detect the target, which seriously affects the efficiency of reconnaissance. By analyzing the imaging principle and design method of the multi-spectral infrared optical system, the reflective structure is selected to design an off-axis three-mirror infrared optical system with a focal length of 50 mm and a full field of view of 12°×10°, which can be used for short-wave infrared imaging, medium-wave infrared imaging and long-wave infrared imaging. Meanwhile, the MTF value of the system is close to or reaches diffraction at the Nyquist frequency of 30 lp/mm in each spectral and each field of view, and its RMS value of the dispersion spot is less than the pixel size of the detector, and all indexes meet the requirements of imaging quality. This discovery expands the detection spectral range of the UAV system and improves the reconnaissance capability of the system.

The cold atom interference experiment involves complex laser modulation technology. It’s important to provide a suitable RF drive signal for the modulator. The phase noise of the signal directly affects the sensitivity of the cold atom interferometer. In order to meet the needs, a low phase noise signal source capable of generating 6.834 GHz signal is introduced in this paper. The signal source adopts two-stage phase-locked loop frequency synthesis technology to multiply the 10 MHz signal output by the constant temperature crystal oscillator to 100 MHz, and then multiply the 100 MHz signal to 7 GHz. Finally the 7 GHz signal is mixed with a 166 MHz signal generated by the external DDS chip. A 6.834 GHz signal source is realized. The test result of phase noise is-61.9 dBc/Hz@1 Hz. It meets the needs of cold atom interference experiments.

Quartz glass is one of the key optical components in the precision optical systems such as UV lithography and laser nuclear technology. The chemical polishing method is proposed to solve the defect issues of the surface or sub-surface damages and etch pits which easily occur in the quartz glass during the machining process. The chemical polishing principle and the working process of quartz glass slices are introduced in this paper. The orthogonal experiment method is used to optimize the chemical polishing process parameters for the quartz glass slices, and the effects of main ingredients of chemical polishing fluid, temperature of polishing fluid and polishing time on the surface roughness are analyzed. The results show that the surface average roughness and the highest transmittanceof quartz glass slices, polished by the mixed solution of NH4HF2 and additive glycerol, can reach about 100 nm and 89%, respectively.