A driving generator to drive multiple infrared Quantum Cascade Lasers (QCLs) was designed under an experimental validation for detection of mixed infrared gases. To avoid the cross-influence of the driving currents among multi-channels, the Time Division Multiplexing (TDM) method integrating with the fast analog proportional-integral (PI) control theory was used to adjust the regulating output currents in each branch independently. A control method combined pulse frequency modulation (PFM) and Pulse Width Modulation (PWM) was proposed to improve the characteristics of the operating frequency and the pulse width of driving pulse and to guarantee the lasers of each branch to operate in an optimum condition. By utilizing the illustrated driver, a driving test was performed on four QCLs with center wavelengths at 4.8, 7.49, 7.71 and 10.7 μm respectively fabricated by Institute of Semiconductor, Chinese Academy of Sciences. Experimental results demonstrate that the stability of driving current is 4.62×10-6 and the linearity of power system is 0.029 1% during a long-term operation (220 h), which meets the requirements of driving multiple QCLs and provides a strong guarantee for mixed infrared gas detection.

To meet the requirements of large-scale equipment for large scope precise measurement in advanced manufacturing industry, a light probe based large Field of View(FOV) 3D vision measurement system based on the binocular stereo vision principle was proposed. The spatial distribution mode of the character points on the light probe was determined according to the invariants of clockwise direction and colinearity under a perspective projection, by which the character points were recognized and the coordinates of the probe were calculated. By using a 3D measurement model based on perspective projection and the homogeneous coordinates of a binocular stereo vision sensor, the structure parameters of the binocular stereo vision sensor were calibrated through linearly solving the essential matrix E, further following with nonlinear optimization by freely moving a 1D target with known precise length. A true large FOV 3D vision measurement system was constructed, which consists of a light probe, a binocular 3D vision measurement system, a portable tripod, a 1D target and a set of measurement software. The real experiment to measure the dense 3D data was performed on a robot body surface in a field of 7 m×4.7 m and the accuracy better than 0.2 mm was achieved. These results show that the designed system has improvement in the light probe structure, recognition methods of luminous points and the system calibration methods.

The engineering suitability of typical linear heat detectors for early fire detection in cable tunnels was explored. In cable tunnel environments, an experimental platform was established based on exiting linear temperature fire detectors, including temperature sensing cables, fiber Raman detectors and fiber grating detectors, then two different fire experimental scenes were set up, and the temperature variation and fire detector response time were recorded in the processing of fire simulation. The experiments show that only directly contacting heat detectors can give an alarm response and other non-contact detectors are unable to respond the small-scale fire because a small amount of thermal radiation has no flame for the beginning of the fire. Furthermore, for massive fire, the alarm response time of both fiber Raman sensing heat fire detector and fiber grating sensing fire detector is about 30 s, the temperature over the fire recorded by a thermocouple is less than 35 ℃ in 60 s, while the temperature rising is more than 5 ℃/min within the range of ±3 m. According to the technical characteristics of the three detectors, the engineering suitability of the typical linear heat detectors in cable tunnels early fire detection is analyzed, which can provide technical support for establishing the design specifications of fire alarm.

On the basis of the inverse Monte Carlo algorithm, a low coherent interferomtric measuring method for the glucose concentration was proposed, and the correlation of the glucose concentration and optical parameters was explored theoretically. The feasibility of reconstruction of absorption and scattering coefficients simultaniously according to interferometric curves was analyzed with the Monte Carlo method. By taking fat suspension Intralipid as a turbid sample, the effect of glucose concentration change on the interferometric signals of depth correlation was researched. Then the scattering and absorption coefficients of the sample were extracted separately and the solution glucose concentration was obtained by the variation from scattering and absorption coefficients indirectly. Results of low coherent measurement was proved by a double integral sphere system and the invariable systematic errors was corrected. With proper correction, it shows that the maximum relative errors are 2.52% and 3.11% for scattering and absorption coefficients respectively. The results demonstrate the strong correlation between glucose and interferomic line as well as scattering coefficient and prove that the scattering and absorption coefficients can be measured in a non-forward scattering.

The effects of preparative parameters such as substrate temperature, ion beam voltage, ion beam current and oxygen flow on the refractive index and stress of a SiO2 thin film were systematically studied by using the orthogonal experiment design method. The transmittance spectrum of SiO2 thin film was measured by spectrophotometers, and its reflective ellipsometric characteristics were measured by an elliptical polarization instrument. Then, the refractive index and stress of the thin film were obtained by the multiple wavelength curve-fitting method and the elastic deformation of a thin film-substrate system, respectively. The experimental results show that the refractive indexes of SiO2 thin film affected by preparative parameters with the weights from high to low are in a sequence of oxygen flow, substrate temperature, ion beam current and ion beam voltage and the confidence probability of effects of the first three refractive indexes is 87.03%, 71.98% and 69.53%, respectively. Moreover, the stresses of SiO2 thin film affected by preparative parameters with the weights from high to low are in a sequence of substrate temperature, ion beam current, ion beam voltage and oxygen flow and the confidence probability of effects of the first three stresses is 95.62%, 48.49% and 37.88%, respectively. It suggests that higher oxygen flows, lower substrate temperatures and lower ion beam voltages should be selected for preparing SiO2 thin films with low refractive indexes and lower substrate temperatures, and higher oxygen flows for preparing SiO2 thin films with low stresses.

As the resolution, accuracy and other parameters of the phase-shifting unit in a phase-shifting lateral shearing interferometer will directly affect the test accuracy of wavefront aberration for a lithograph objective, a phase-shifting component with macro and micro compound motion modes was designed based on the principle of phase-shifting test for wavefront aberration. It could achieve a phase-shifting resolution of 3 nm in a 25 mm travel range. Moreover, the mathematical relation of initial parameters was analyzed for the micro unit with a flexure hinge composited four linkage structure. The stiffness and weak interfacial stress of the flexure hinge were calculated, and a design example was given. The finite element analysis method was used to simulate the relation between phase-shifting value and the output of piezoelectric ceramic (PZT) and to analyze the phase-shifting accuracy. The results show that the PZT can actuate one-dimensional phase-shifting motion in 0.1 mm-1 nm within its output range, and the theoretical accuracy is better than 3.5 nm. The open-loop calibration test shows that the actual accuracy of micro-motion unit in the phase-shifting component is better than 5 nm.

To develop a location system for small agricultural machines in the Chinese south, a laser emitting and tracking system based on double laser-emitter and three-point location was designed based on the laser three-point theory. The theory of double-laser-emitter and three-point location was introduced; then, the laser emitting and tracking system was designed, including the designs of working mechanism and controlling circuit for the laser emitter. The Kalman filtering algorithm based on a incremental PID with quick correction functions was proposed to estimate the moving targets in laser tracking and to compensate the errors in changing speeds and turning corner of the machine. A series experiments were performed. The experiment results show that the targets of laser receiver can be tracked dynamically by both laser emitters in real-time, and the maximum deviation is ±2.8 cm, which meets the location requirement of an autotracked-driving farming machine for an accuracy of ±4 cm within 200 m. For the reliable and effective algorithm, it can lay a foundation for unmanned driving of farmland machines and data collection in farming fields.

On the basis of an optical modulation model and a normalized correlation model, an objective evaluation method for the resolution of a low-light-level image intensifier was explored by using the dual criterion of template matching and optical modulation as core algorithms. A set of resolution objective evaluation system for the low-light-level image intensifier was constructed. In the method, the optical modulation model and the normalized correlation model are independent of each other and also cannot be separated from each other. Firstly, the template matching coefficients in the normalized correlation model were used to performed the location and preliminary evaluation , then the optical modulation model was taken to do a quantitative analysis. The image processing scheme in first qualitative analysis then the quantitative analysis improves the accuracy and repeatability of resolution test for the low-light-level image intensifier, implements the objective and accurate evaluation for low-light-level image intensifier and avoids the shortcomings of visual observation method which is influenced by human's subjective factors. The experiments demonstrate that the resolution of the image captured by CCD has good consistency with that of visual observation result by human eyes. The method is also suitable for the evaluation of resolutions of other visible light imaging systems.

Parallel plate capacitor is a core mechanism in a Micro-mechanical-electrical (MEMS) sensing device. As the surface roughness of an electrode has obviously impact on the space electric field when the distance between the electrodes is shortened in the capacitance detection, this paper explores the effect of the surface roughness of electrode on the performance of the parallel plate capacitor. A parallel plate capacitor model with a single roughness electrode was established, and then the finite element method was used to study the effect of the surface roughness on the detecting capacitance. Based on increasing rough surface to enhance the memory electric charge capability, the formula for parallel plate capacitor with rough surface was corrcted. Finally, the Atomic Force Microscopy(AFM) was used to describe the samples with different surface roughnesses. Experiments and simulation results indicate that the surface roughness has a obviously effect on the detecting capacitance. Increasing the surface roughness of an electrode and decreasing the distance between the electrodes can improve the detecting capacitance greatly. When the surface roughness of an electrode increases from 0.063 nm to 60 nm, the detecting capacitance value grows by 9.0 percent. The result shows that increasing the surface roughnesses of electrodes can improve the sensitivity of MEMS devices.

A highly sensitive and high selective sensor which can detect drug molecules in a liquid environment was explored. By combining the Molecular Imprinting Technique (MIT), Film Bulk Acoustic Resonator (FBAR) and the micro fluidic technology, the new sensor was constructed. The preparing principle of adsorption layer materials for FBAR by the MIT was introduced and the Imprinting material with special adsorption properties for the special drug was developed by the MIT. Then, the material was coated on a FBAR as a adsorption layer based on a Micro-electrical-mechanical System and the micro fluidic technology. According to the piezoelectric equation and dynamic equation, the input impedance formulas of the equivalent circuit and electric port for the FBAR piezoelectric vibrator were derived. Furthermore, a model of the micro fluidic system with the MIT plus FBAR-based drug detection sensor was designed and a detailed theoretical derivation and analysis was performed. Finally, the sensor model was simulated using practical operating parameters. The results show that the quality of the mass loudness per unit area of the sensor is up to 0.8 pg/Hz·cm2, which is much higher than the quality of a conventional Quartz Crystal Microbalance (QCM) sensor. This advantage can greatly enhance the performance of the sensor for drug detection in the liquid environment.

A driving power supply was designed for the two-degree-of-freedom (2-DOF) ultrasonic motor with a single plate-shaped vibrator. According to the heavy current burden and low power utilization of the driving power supply, a method for optimizing the driving performance was proposed. The power supply used a voltage-controlled oscillator to generate a small sine signal, and applied a high-voltage and power amplifier to produce a higher driving voltage and a larger driving current. According to equivalent circuit and operating characteristics of piezoelectric vibrator, the ultrasonic motor was connected with an appropriate inductor in parallel to optimize the driving performance of the driving power supply. A prototype of driving power supply was made and its output performance and optimization effect were tested in an experiment. Experimental results show that the control variables of ultrasonic motor in power supply-driving voltage and frequency can be adjusted independently and continuously, and an ideal waveform of sine signal is obtained. When the power supply is loaded by the 2-DOF ultrasonic motor in frequencies of 49.127 kHz and 49.756 kHz, the maximum output voltages are 176.0 V and 171.2 V. When the ultrasonic motor is connected with a 1.442 mH inductor in parallel, driving current can be reduced to 7.27% and 7.41%, and the power factor can reach 0.968 and 0.955 0, respectively. The power supply designed meets the driving requirements of the ultrasonic motor, and optimization method on reducing current burden and improving power utilization has good results.

A transferrable arithmetic of the position-disturbing value for a stable platform was proposed on the basis of the detailed study of the stable principle of the dynamic platform and its configuration characteristics . The principle of the transferrable arithmetic is as fellows: the coherence is built between the middle inertia and the platform firstly, and then the relation conversion matrix is established between the vehicular inertia and the middle inertia. The value of the middle inertia is deduced by the conversion matrix and vehicular inertia. By above, the middle inertia can be replaced with a “fictitious inertia” constructed by the relation conversion matrix and the vehicular inertia. This “fictitious inertia” value is the position-disturbing value of the dynamic platform. Experimental results show that this transferrable arithmetic has high precision, and its max error is 0.0561°. The dynamic platform stabilization is realized by this arithmetic with the ready inertia on the basis of the unadjusted vehicular inertia. It has advantages not only the work efficiency but also the performance-price ratio. The method has been already successfully used in several vehicular dynamic platforms.

A Micro-electro-mechanical System(MEMS) fiber-optic magnetic sensor based on a MEMS torsional mirror was proposed and a dual-fiber collimator was used for measuring the tiny angle of torsional mirror. The MEMS fiber-optic magnetic senor is consists of a MEMS torsional mirror, a magnetic film and a dual fiber collimator. The mechanisms of magnetic sensing and optical detection of the device were described, and the design and optimization of the device were discussed. A prototype for the MEMS magnetic sensor with a volume of 3.7 mm×2.7 mm×0.5 mm was fabricated successfully by MEMS technologies. The measured output values of the magnetic sensor are consistent with theoretical values. Experimental results indicate that the sensitivity of the magnetic sensor is 0.65 dB/mT. and its minimum resolution magnetic field is 167 nT. The MEMS fiber-optic magnetic sensor combines fiber-optic measurement and compact MEMS structure, and it has advantages of compact construct and simple fabrication processes. Moreover, it can operate without current excitation.

Ion beam etching technology was researched for fabricating blaze concave gratings, and a partitioned space calculated method was proposed to calculate the diffraction efficiency of a blaze concave grating. The method can calculate the diffraction efficiency quantitatively without experimentation by confirming an incidence angle. The blazed angle of the blaze concave grating with a given wavelength was calculated based on the partitioned space calculated method, and ion beam etching parameters and a mask parameter were calculated by etching software BLAZING. On the basis of these parameters, a concave grating with a size of 45mm×40mm was fabricated by holographic ion beam etching. The focus blazed angles measured by an Atomic Force Microscope(AFM) are 9.21°,the efficiencies of the gratings measured by a concave grating diffraction efficiency instrument are about 54.8% at 300 nm and 50% at 250 nm. The grating substrate has a smooth blazed groove profile with 1 200 groove/mm. The results are in agreement with that partitioned space calculated well. It means that partitioned space calculated method is suitable for designing and calculating the efficiency of blaze concave gratings and it can direct the fabrication of the blaze concave gratings by the ion beam etching technology.

To count the cells precisely by using a microfluidic cytometer, a three-dimensional focusing microfluidic chip was designed by UV-photolithography based on SU-8 photoresist. The Polydimethyl Siloxane (PDMS) was used to cast doubly to replicate its structure and to shorten the fabrication cycle of microfluidic chip and reduce its costs. First, the PDMS was taken to fabricate a mold with a three-dimensional focusing microfluidic channel by the oblique exposure lithography of SU-8 photoresist and immersion lithography. Then, the PDMS was used to cast firstly for the SU-8 microfluidic to obtain a PDMS negative mode structure. After the negative mode structure was treated, it was casted again and the PDMS microfluidic detection channel which was consistent with that of the original mold was achieved. Finally, the PDMS microfluidic channel with a cover plate was packaged by using a certain packaging method and the microfluidic chip was obtained. Furthermore, the channel focus effect was tested. The test result shows that the focused width of sample flow changes with the different velocity ratios of sheath flow and sample flow. When the ratio of sheath flow and sample flow is 20∶1, the focused width is about 10.4 μm. The results demonstrate that the chip with the microstructure is reliable, and it meets the requirements of further fluid focus testing. The manufactured microfluidic chip by using this process has a short cycle, low costs and a high efficiency.

To realize the flexural support for optical elements in a precision optical instrument, a Cartwheel bi-axial flexural hinge composed by filleted short beams was proposed. The dimensionless design graph method for the design of the spatial flexural hinge was presented. First, the parametric finite element analysis on the Cartwheel bi-axial flexural hinge was performed, and then a polynomial fitting was carried out according to the analysis results to establish the dimensionless design graph for the mechanical characteristics of the flexural hinge, such as rotational stiffness and maximum stress. A practical design by the dimensionless graph method was performed to satisfy the supporting demand of an optical instrument, and the finite element analysis was used to verify it also. Finally, an optical test platform was established, and the rotational stiffness of the design was measured. Obtained results show that the maximum relative error of the rotational stiffness between analysis result, test result and design result is 10.1%. In conclusion, by using the dimensionless design graph as a design tool, a designer can determine the optimal geometry rapidly and correctly of the Cartwheel bi-axial flexural hinge based upon its demands for the stiffness, rotation angle, maximum stress and the weight. This paper can provide reference for the application of the Cartwheel bi-axial flexural hinge in precision optical instruments.

A tri-axis gyroscope with the improved square symmetrical structure consisting of four large proof-masses and four small proof-masses was proposed to realize a monolithic six Degree of Freedom inertial measurement unit. Firstly, the structure and its working principle of the unit were introduced and the required performance was given by detailed design parameters. Then, the modal analysis for in-plane driving mode and out-plane sensing mode was simulated. To increase the mechanical sensitivity for each sense mode, a dual electro-statically tuning method was adopted. In this method, a tuning voltage was first applied on the proof-mass to make the frequency shift between drive mode and yaw mode as small as possible, and the other tuning voltage was applied on a polysilicon electrode plate over the small proof masses to further make the frequency shift among the four modes to be under 30 Hz. By these, the tuning method was verified to be effective. Furthermore, the quality factors for driving and sensing resonant modes were simulated through analysis of the thermoelastic damping mechanism. The analysis results indicate that the quality factors in driving and sensing modes are 23816 and 19507, respectively. Finally, a process flow for high aspect was designed combining the surface and bulk micromachining. The design and simulation for this square symmetrical and decoupled structure demonstrate that the mode matching and quality factor can meet the design requirements for the tri-axis gyroscope.

A portable minimally-invasive human blood glucose detection instrument was designed by a microfluidic chip based interstitial fluid transdermal extraction system. On the basis of the minimally-invasive method, the instrument utilizes a vacuum pressure to extract the interstitial fluid from human body, adopts the Surface Plasmon Resonance (SPR) technology to measure the glucose concentration of the interstitial fluid, and then predicts the blood glucose concentration with the glucose concentration of the interstitial fluid. Moreover, by immobilizing D-galactose/D-glucose Binding Protein (D-GGBP) which has a specific adsorption function to the glucose, it preprocesses the surface of the SPR sensor and realizes the specific adsorption of glucose molecules. In the experiment, the glucose solution with different concentrations was prepared and detected. According to the experimental results, the curve illustrating the relationship between glucose concentration and refractive index was obtained. In addition, the glucose solution and interstitial fluid were measured by using minimally-invasive blood glucose detection instrument designed by our research group. The test result was compared with that measured by a glucose meter. The experimental result indicates that the glucose detection resolution could reach 0.625 mg/dL using the SPR sensor decorated with the GGBP. It shows a good linearity when the glucose concentration ranges from 0.625 mg/dL to 5 mg/dL. The experiment verifies the feasibility of the instrument, and demonstrates that the SPR technology with immobilized GGBP will obtain a wide application in the field of minimally-invasive glucose measurement.

To perform precise measurement on mega gears, the principle of mega-gear in-site measurement system with laser tracker and corresponding key techniques are originally introduced. Considering the mega-gear in-site measurement system with laser tracker integrates the large-scale measurement capacity of laser tracker and the high accuracy of coordinate measuring machine, the laser tracker is used to establish the coordinate system both for mega gears and the Three-coordinate Measurement Platform (TMP), thereby, their coordinate systems are associated by the coordinate system of the laser tracker, and the module and algorithm of coordinate system fitting are proposed simultaneously. Meanwhile, a novel module of attitude adjustment of TMP is established, which can be accomplished by the system of attitude adjustment to ensure that the position relationship between TMP and the axis of gear can meet the measuring requirements. The results illustrate that the principle of mega-gear in-site measurement system with laser tracker is correct and feasible, and in consequence it can meet the six grade precision measurement requirements of mega gears.

As existing objective evaluation index for color image fusion is inconsistent with the human vision perception, a non-reference index based on quaternion convolution was proposed. First, a color image was modeled in a holistic manner, in which the color information of the color image was considered fully as a whole. Then, the quaternion-valued edge detection template and the color image were used to do a convolution operation and to get the detailed color information. Furthermore, the image definition and useful information from the fusion were measured and they were given by weight modes. Finally, a set of quantitative computations for the fusion images were performed and objective evaluation results were given. The experimental results show that the proposed method can utilize the color information and other detail information obtained by human vision. It works better than the traditional methods, and shows a better stability in the color image fusion evaluation. The evaluation results of proposed method are consistent with the human vision perception, and fulfill the needs of objective color image fusion.

Workspace measurement and positioning system works at multi-station in a synergistic effect and how to deploy the stations has a significant impact on the measurement range, accuracy, and the use-cost. This paper explores the effect of the geometry of station distribution on position errors based on the relation between the wMPS network deployment and position errors. Firstly, a positioning error model was established, the station one-way communication constraints were analyzed, and the improvement of increasing the number of stations on measurement accuracy was analyzed. Then, typical deployments of two to four stations were designed and the error distribution characteristics were studied. Finally, the error characteristics of typical deployments were verified based on the latest prototype in a lab. Experiment results show that O~~4 typical deployment has the highest accuracy, the measuring accuracy from the L~~3 typical deployment is higher by 40% than that of the I~~2 deployment. Moreover, O~~4 typical deployment has 20% improvement on measuring accuracy than that of the L~~3 typical deployment. The research work provides an effective theoretical support for global measurement network optimization.

To improve the performance of person tracking for a mobile robot in complex environments, an adaptive kernel function based Mean Shift algorithm was proposed by using a coarse to fine localization mechanism. In the outer layer, a Radio Frequency Identification Device (RFID) was adopted to detect the person with an ID tag to determine the Region of Interest (ROI) coarsely. In the inner layer, the ROI of a disparity image was processed to estimate an initial searching window. Then, the adaptive kernel based Mean Shift algorithm was applied to location of the person precisely in the left image from a stereo camera. The adaptive kernel function was combined with the regional feature of person and the Epanechnikov function, which can reduce the effect of the background pixel on the target’s color probability distribution. Compared with the traditional Mean Shift algorithm, the presented algorithm can track the target successfully when the background has the same color. Furthermore, the searching area is narrowed by the RFID, so that the computational cost is reduced. The average computing time is 62.11 ms/frame, which satisfies the requirements of real-time target tracking. The experimental results indicate that the proposed tracking method can complete the target tracking in a background with the same color, short-term occlusion, fast moving, and a sudden turn for a mobile robot.

On the basis of properties of magnetic resonance images for the prostate, an active shape image segmentation method making use of adaptive texture distribution was introduced to segment a prostate magnetic resonance image. Firstly, a prostate region of interest was determined through image classification and image fitting, and several shape parameters were estimated and used in the segmentation. Then, multi-features were fused to build a texture coincidence measure. In order to improve the searching and matching ability of an active shape, the active shape was divided into two portions, the texture distribution shape and the supplementary shape. In search, the estimated parameters were used to optimize the initial estimation of the active shape searching and adjust the iterative process based on the texture distribution shape and the supplementary shape. Experimental results indicate that the Hausdorff Distance is 6.00 pixels between the true prostate contour and that extracted by the proposed method and the segmentation accuracy of the method is 93%. The proposed method can modify the active shape effectively, and can automatically segment the prostate magnetic resonance images with high enough accuracy.

An automatic extracting method for the image of inner contour of a vessel wall from intracoronary Optical Coherence Tomography (OCT) was investigated to improve the robustness and computing efficiency of traditional algorithms. Firstly, the correlation between OCT images in volume data was used to remove the effect of the signal caused by a catheter on the edge extraction. Then, the ray shooting method was taken to estimate the center of vessel, and the center was chosen as the pole to do a polar transform on the OCT image. Furthermore, the transformed image was partitioned into several parts, these parts were denoised separately by different thresholds and the upper edge was extracted from the denoised image. Finally, the inner contour of the vessel wall was obtained by an inverse polar transform. Experimental results on more than 400 images indicate that the inner contours of vessel walls of intracoronary OCT images can be detected accurately, and the average processing time for each image is about 1.2 s.In conclusion, the method can deal with OCT images with vascular bifurcation and ones influenced by catheter largely at higher speeds, so it has advantages in robustness and computational efficiency.

For the peculiar Smear noise occurred in the area CCD imaging, a method to extract the Smear from the interline transfer CCD dark pixel region was proposed by taking an interline transfer area CCD for an example. The working principle of the area CCD was introduced, the creation reason of the Smear was analyzed and the composition of the Smear noise was discussed. Then,the KTC noise and the dark current noise of the Smear were filtered by using the Correction Double Sample(CDS) and the dumb pixel correction. The average filter method based on a middle value was proposed to eliminate the light shot noise. Finally, the Smear was eliminated from a original image by the subtract method, and the Smear image region was compensated by the triangle-based interpolation. A real-time hardware processing system was designed by using Field Programming Gate Array +Digital Signal Processor (FPGA+DSP) as the core processor. When the camera is working at the fastest working mode 3 frame/s, the Smear can be extract and filtered in 1.265 ms and the gradation variance of the image area has decreased to 95.34% by eliminating the Smear. With the imaging experiment verification, the system has a high integration and real-time ability and the Smear noise is eliminated completely.

Several local region partition methods for Speeded Up Robust Features (SURF) descriptors were researched to reduce their dimensions and increase the matching speeds and robustnesses of SURF based matching algorithms. With reference to existing local region partition methods of Scale Invariant Feature Transform (SIFT) and SURF descriptors, the local regions were divided into grids(original SURF), triangles, and sectors. First, the influences of scale change and rotation of the image on the matching performance of SURF descriptors were analyzed. Then, a method to construct the SURF descriptors with local region partitions in triangles and sectors were proposed, and matching experiments were performed. The SURF descriptors with different local region partitions were compared. The experiment results show that the performance of the sector partition based SURF descriptor is better than those of triangle partition and grid partition (original SURF) based SURF descriptors. The performance of SURF descriptors with 6-sector partition, 8-sector partition, 12-sector partition and triangle partition is better than that of original one, and the dimensions of these new descriptors are 40, 32 , 16 and 32 lower than that of original SURF (64 dimensions).

A robust non-rigid registration approach with a t-distribution Mixture Model(TMM) was proposed because point sets as Gaussian mixture model is vulnerable to the outliers and the data with longer than normal tails. The Gaussian mixture model was extended to the student′s-t Mixture Model by full data definition in the Expectation Maximazation(EM) frame, then the closed solutions of the parameter set of the t-distribution mixture model were solved by re-parameterization of the t-distribution mixture model in the EM algorithm. The priori-weight of each float point was calculated in EM framework to reduce the effects of outliers and the data with longer than normal tails on the matching results. The degree of freedom of each float point in the t-distribution mixture model was calculated to change the probability density distribution, improve the robustness of the algorithm and avoid the effect of estimating the outlier level of point sets that may bring additional errors. The conditional expectation function in t-distribution mixture model was added a regular item of point set, so that the points have a feature of Coherent Point Drift(CPD). The simulation data show that the error from the TMM-CPD is only one tenth of that from comparison algorithms. When the point sets are approximate ellipse shape, tubular and three dimensions, the registration errors of TMM-CPD are only 42.0%, 80.1% and77.5% of those using comparison algorithms, respectively. The experiments show that this non-rigid registration approach using the t-distribution mixture model has features of high-accuracy, good robustness compared to other point set registration algorithms for point sets containing outliers and data with longer than normal tails.

To ensure the angle measuring accuracy of a shipboard theodolite, and to compensate the errors of the station position and ship-swaying measurement, the effect of attitude data on angle measuring errors was analyzed and a ship-swaying error model was established. On the basis of the model, a scheme for calibrating and measuring the equipment was given, and then a shipboard stations position correcting algorithm and data post-processing method were proposed. First, the ship-swaying error model was build, and the formulas for the effects of measuring errors from the ship′s head, pitch and the roll on measured azimuth and elevation angles of the equipment were given. Combined with the equipment indicators and mission requirements, a scheme for calibrating and measuring was established and the measured angle ranges to ensure the accuracy of the equipment were given. Furthermore, a measuring method of station position and its correcting method were proposed when the equipment works at special working environments. Finally, a data post-processing method for shipboard equipment was explained. Experimental results show that the Root Mean Square(RMS) of angle measuring errors of the equipment are less than 57″ for the azimuth, less than 34" for the pitch when the shipboard accurate attitude measurement system has angle measuring errors at the head of 1.2', pitch of 24″ and the roll of 24″, respectively. The result has basically reached the optimal theoretical accuracy under this measurement conditions.

Uniform linear motion blur often occurs in the camera imaging, which results in image degradation seriously. Therefore, this paper proposes a method to identify the motion blur extent accurately. Since a motion blur angle could be identified accurately by Radon transform and the motion blur direction could be rotated to the horizontal, only the horizontal motion blurred images are needed to be researched. For a motion blurred image without noises, the Fourier transform is used to transform it to frequency domain, the motion blur length is estimated by the BP neural network and the input of BP network is the sum of the amplitudes in the central region of the spectrum. For a blurred image with noises, the bispectrum transform is applied, the motion blur length is estimated by BP neural network and the input of BP network is the maximum value of each column of bispectrum. The simulation experiments show that the proposed method in this paper is correct and efficient. When the SNR is larger or equal to 23 dB in noisy images, the mean error of blur parameter identification by proposed method is below 5%, which is superior to that of conventional schemes.

The image tailing generated by an Extreme Ultraviolet imager was analyzed and how to remove the image tailing phenomena was proposed. The effects of each component change of coordinate calculation formula on the image were analyzed, and it shows that the superposed signal and Micro Channel Plates (MCP) feedback are the most influential factors of coordinate components in the system. After image data analysis, both cases of superposition of signal, peak accumulation and tail accumulation, were excluded. The calculation on change of quantity of charge at the most serious tailing place shows that the MCP feedback may be the reason of image tailing. However, the amplitude of high voltage applying on the MCP is one of the important factors affecting feedback occurs. The images sampled at 2 950 V and 2 800 V were validated and it indicates that the tailing is sparked by the MCP feedback. Moreover, solutions were given, that are to reduce rudimental gas in the MCP by baking and to clean the MCP by electron beams regularly.

To estimate Doppler frequency and Direction-of-Arrival (DOA) accurately and efficiently, a joint spectrum estimation method based on Weighted Subspace Fitting (WSF) algorithm and Quantum-behaved Particle Swarm Optimization (QPSO) algorithm, namely WSF-QPSO, was presented. First, an extended observability matrix containing the information of Dopplers and DOAs was constructed by using a state-space model, and the joint spectrum function was fitted by using WSF algorithm. Then, the joint parameter estimation was converted to multidimensional nonlinear function optimization. Finally, the Dopplers and DOAs were estimated by optimizing the joint spectrum function using QPSO algorithm. Experimental results indicate that the Root Medium Square Errors (RMSEs) of Dopplers and DOAs estimated from the WSF-QPSO method are 0.007 5 rad and 0.25°, respectively when SNR is 0 dB. The proposed method can get high resolution and robust parameter estimation with less control terms, and the parameters are paired automatically. In addition, the WSF-QPSO method can obtain acceptable estimation results even under the condition of low SNR or small snapshot number in comparison with the joint spectrum estimation methods based on subspace decomposition.

An algorithm combining Center Surround Extremas(CenSurE) and spatial-temporal information was proposed to improve the detection speed and detection integrity for a moving object in dynamic scenes. Firstly, in terms of the rapidity and accuracy of CenSurE feature point extraction, an inter-frame image registration method for the motion sequence base on CenSurE and a homography transformation model was proposed to compensate the translation factors, rotation angles and scaling coefficients caused by a moving camera. Then, the foreground mask was generated according to the temporal inter-frame difference of the registrated background in a time domain. Moreover, a dynamical updating background was established with spatial information of the foreground mask and a complete moving object was extracted through background subtraction and a self-adaptive threshold segmentation method based on probability and statics. Finally, the experiments on several standard video sequences were tested to evaluate the performance of this algorithm. Experiment results show that the operating speed of this proposed algorithm can be 15 frame/s, meanwhile, complete moving objects can be obtained with high detecton speeds. The proposed algorithm basically meets the demands of moving object detection in dynamic scenes for speeds, noise resistances, light adaptability, object integrity and so on.

To segment fuzzy and touching cell images accurately, an image segmentation algorithm based on graph theory and morphological mathematics was proposed according to the characteristics of medical cell images. With proposed algorithm, the images were smoothed and sharpened, firstly. Then, the improved Minimum Spanning Tree (MST) algorithm was used to segment the cell images, in which the cell size and shape information were added into MST graph to avoid the over-segmentation. Furthermore, the adherent cells were split by combining the distance mapping and the skeleton information in morphological mathematics to solve the problem of cell adhesion in the binary image. Different from the traditional watershed algorithms, the split algorithm has no repeat operation. By experiments, it shows that the proposed algorithm can segment fuzzy and touching cell images well and can obtain desired results.