A self-developed large aperture near-infrared phase-shifting flat interferometer was used to perform an absolute test for a large aperture SiC flat mirror in an oblique incidence condition. Firstly, the cavity wavefront data were obtained in a standard Fizeau configuration, and then the test flat was put into the cavity in its normal axis with an angle α to the optical axis of the interferometer to obtain the second wavefront data set. After processing the two data sets, the absolute vertical profile across the center of SiC mirror was also obtained.Finally, several absolute vertical profiles of the Φ630 mm SiC mirror were measured and the results indicate that PV and RMS of the center absolute vertical profiles are 0.061λ and 0.014λ, respectively. The experimental results show that this measurement device can achieve the absolute tests of vertical direction of flat mirrors with the effective apertures larger than that of the interferometer, and it is especially applicable for the absolute tests of optical sur-faces coated with high-reflection films or metal surfaces.

According to the configurations of optical system in a three line array mapping camera, a new optical system integrated with excellent features by Petzval and symmtry was designed based on a telecentric path to meet the requirement of stereoscopic mapping.The system shows a good imaging quality, and the MTF value of each lens is close to the diffraction limit at the Nyquist frequency of 77 lp/mm. Furthermore, the distortion of each optical system is less than 2.0×10-5 for the straight-sight camera and 2.4×10-5 for forward-sight and backward-sight cameras. Sensitivity analysis and inverse sensitivity analysis were used to compute the effect of each tolerance parameter on the optical system performance, and the proper tolerance requirements and compensators were determined.On the basis of tolerance given, the optical system with high image quality and low distortion was fabricated and assembled. The tested result shows that the MTF of each optical system can reach 0.451 at the Nyquist frequency of 77 lp/mm, which meets the requirements for static MTF more than 0.2 and relative disortion less than 3×104 and proves that tolerancing analysis methods are correct and feasible.

As the iodine laser is pumped by chemical energies at present, it shows a larger cubage, shorter running time and produces toxic materials easily. These disadvantages limit the application of the iodine laser. To meet the requirements of the miniature, pollution-free and high power iodine laser in practical applications, a new idea for iodine laser production was proposed in this paper. Firstly, a plasma generation model was established to produce a 1315 nm iodine laser stimulated emission.To obtain the optimum parameter range of an experiment in our experiment conditions, the power density of pump light, ionization electron density and the free electron temperature were calculated. Then in the optimum parameter range calculated, the iodine plasma generated by a single longitudinal mode second-harmonic Nd∶YAG laser (532 nm) was detected by a spectrometer to obtain its fluorescence emission spectrum. Finally, the fit result for the signal obtained in the experiment shows that the signal peak is located at 1313 nm with a bandwidth of 1.182 nm. The results reported here demonstrates the possibility that the iodine plasma is acted as laser media to produce stimulated emission.

For the stray light from an Extreme Ultraviolet (EUV) solar telescope in the ground experiment, a measurement system for the Point Source Transmittance (PST) was established based on the parallel light introduced by the rotary scanning mirror at various angles and the PST curves were applied to assess the stray light level. By analyzing the structure of the telescope and the measured PST curves, the first order stray light under a small angle range was proved to be the main source of stray light.To eliminate the stray light, the principles of the baffles were presented in details and the simulation modeling of the telescope with baffles was created by Tracepro software. The eliminating ability of the baffle was calculated and tested at different Bidirectional Reflectance Distribution Function (BRDF) values. The test results show that the baffles can block most of the first order stray lights,the intensity of stray light has reduced 3 orders of magnitude,and the corresponding veiling glare index is less than 3.65%. With the help of baffles, the EUV telescope performs near the diffractionlimited condition at UV band, and all the requests for the imaging test can be met.

A real ray tracing method was used to simplify the optical design of a Gaussian beam shaping system. The principle of shaping Gaussian beam was studied theoretically and the Flattened Lorentzian(FL) function was chosen as the distribution expression of the flattened beam. The relationship of coordinate transformation of arbitrary rays in an incident plane and an image plane was deduced based on the law of energy conservation. Then, according to the characteristics of this system, Zemax Programming Language (ZPL) was used to compile ZPL macro orders to calculate the coordinate transformation, and the real ray tracing method was adopted to design the optical Gaussian beam shaping system. Finally, the aspheric lens system was processed by single point diamond turning techniques and its shaping ability was tested by the optical analyzing software. Testing results indicate that the system can achieve the conversion from the Gaussian beam to the flattened beam, and the uniformity of flattened beam is 87.1%. The method is not only simple but aslo practical and has a significant engineering application value.

A high efficient, compact, Refractive/Total internal reflection/Reflective/Refractive (RIXR) collimating system was designed and optimized to collect a wide angle range light to emit the collimating beam with a small divergence angle. Firstly, the surface of the initial structure of the system was calculated according to the edge-ray principle and Simultaneous Multiple Surfaces (SMS) method based on the non-imaging optics theory. Then, the initial structure was optimized by using Zemax optical design software. Finally, the best LED collimating system was successfully designed. According to the results of the design, the semi-diameter of the RIXR collimating system is 20 mm, aspect ratio is 0.25 and its collection angle is 200°. These results indicate that the system can completely collect the light emitted by LED sources. As the root mean square radius of a spot diagram for the edge field (± 3.178°)obtained by the reverse ray tracing is less than 2.1 μm, the edge field shows an excellent focusing effect.Furthermore, when the light absorption loss of the material and reflection loss of the interface are fully considered, the light energy utilization rate is as high as 83.48%. Experiments demonstrate that the RIXR-type collimating system has advantages in a compact structure, high light energy utilization ratio and easy to use.

To overcome the limitations of traditional deformation measurement methods and complete the comprehensive measurements of the general deformation and the various point displacements within a measurement scope in any time, the Digital Speckle Correlation Method(DSCM) was used in accurate deformation measurement. Firstly, by comparison with traditional methods, it summarized the advantages of the method in noncontact, high precision, simple and flexible operation, overall measuring and so on. And then the principles of the DSCM were elaborated on the five aspects, the relevant mathematical models were established and the corresponding measurement and analysis software XJTUDIC was developed. By using this software with the relevant hardware devices, a tensile deformation process was tested and obtained results were compared with that of traditional measurement method using an extensometer. Obtained results indicate that the error is within 0.2%, which fully validates that the method and corresponding software and hardware are reliability and feasibility and it provides an effective and reliable tool for the precise and comprehensive deformation measurement.

This paper researches how to realize the Butterworth type Frequency Selective Surface(FSS) from a symmetric double screen FSS and to obtain the “rectangular” filter characteristics with “flat top” and “steep cut-off”. Combined with the plane wave expansion and mutual admittance methods, the aperture array admittance and the mutual admittance between the arrays were given. Based on the method above, the effects of the FSS screen and the dielectrics media between the FSS screens on the Butterworth FSS were analyzed numerically by taking a common “Y” aperture for an example. Then, a single screen “Y” type FSS was fabricated and pasted on the both sides of media accurately by plating a copper film with thickness of 15 μm on a 500 mm×500 mm polyimide substrate. The free space method was used to test the transmission characteristics, and the simulation results are in good agreement with test ones. Furthermore, when the electrical thickness of the dielectric media was set to be in 2.15 mm, “Y” unit cell period in 7.2 mm×6.235 2 mm, arm length in 3.6 mm, and the arm width in 0.8 mm, the mutual admittance of dual screen FSS is equal to the real part of admittance for a single-screen FSS. As a result, the Butterworth type FSS is obtained successfully. Experiments show that the designed Butterworth type FSS has reduced its -3 dB bandwidth from 3.5 GHz to 2.1 GHz when the cell circle is increased by 0.6 mm. Finally, it suggests that the design of Butterworth type FSS should be to adapt a no-load cell unit with a circle less than 0.4λ on the assumption of resonant sizes, and its radio between gap length and width and the thickness of intermediate electric should be less than 5.5 and about 0.1λ, respectively.

On the basis of the wavelength scanning and spectrum modulation techniques of an ultra-narrow-linewidth semiconductor laser, a kind of multi-component gas online detection system with high precision was designed. A semiconductor laser with ultra-narrow-bandwidth and tunable wavelength was utilized as a light source and a novel series gas cell was also used to compress the horizontal space and increase the optical path of the system. By taking the sawtooth signal to modulate the spectrum of laser and using space division multiplexing (SDM) and time division multiplexing (TDM) techniques, the detection of multi-component gases(CO, CO2, CH4) based on ultra-narrow-linewidth laser characteristics was achieved. The experiments indicate that the relative measurement error is less than 2% and linearity relevance coefficient is 0.99 for the three kinds of gases. Furthermore, the system dynamic response time is less than 10 s by filling a volume of multi-component gas into the gas cell gradually. Finally, the result of long time stability monitoring shows that the maximum concentration fluctuation is less than 0.02%. The system has advantages of high accuracy, good stability and quick response, and is suitable for the fault gas monitoring of power transformers in real time.

According to the characteristics of a catadioptric omnidirectional imaging system in larger fields and higher noise shielding, the catadioptric omnidirectional optical system with a field of view in 360° horizontal azimuth and 45° to 90° elevation and operating in the region of Solar Blind UV was developed to detect UV targets.The key elements of the UV detection technology were overviewed and the design parameters for the catadioptric omnidirectional imaging system in Solar Blind UV(SBUV) were determined.Based on the aberration theory and characters of UV optical system, the processes of optical design including the allocation of parameters, solution of initial structure and the lens splitting were presented. The design results show that the diameter of 80% encicled energy of SBUV panoramic imaging system is 20 μm, which is smaller than the pixel size of ICCD, and the MTF value of the system is higher than 0.7@17lp/mm over all fields. These data satisfy the requirements of the system. Furthermore, experiments are carried out to verity the feasibility of panoramic imaging in corona detection, and the results indicate that the image obtained has good quality.

In order to improve the surface hardness and corrosion resistance of stainless steel, single-pass and multi-pass lap laser cladding coatings of Ni25WC35 alloy powders were fabricated on a 1Cr18Ni9Ti substrate by a 6 kW transverse-flow CO2 laser. The coatings were examined and tested for chemical compositions, microstructure features, phase structures, microhardness, corrosion behaviors and metallographies by the X-ray Diffraction (XRD), Scanning Electronic Microscopy (SEM), X-ray Energy Dispersive Analysis (EDAX), microhardness tester, electrochemistry workstation and optical microscopy (OM), respectively. The results show that some new harden phases are formed in the coating which consists of (Fe, Ni) solid solution and some compounds such as WC, W2C, CrNiFeC, FeW3C, Ni2Si, Fe3Ni3B, and these harden phases improve the hardness and corrosion resistance of the substrate effectively. OM observation displays that the microstructure of clad coating is homogeneous, compactness and forms a good metallurgical bonding layer with the substrate. Micro-hardness measurement exhibits that the surface hardness of treated sample is twice as large as that of substrate, and the highest hardness lies in the second pass cladding zone center of a double-pass coating sample and reaches 650HV. The electrochemical corrosion test in 5.0% NaCl saturated solution indicates that the maximum self-corrosion potential (Ecorr) of clad coating sample changes from 630.90 mV to positive direction, and the lowest corrosion current density (Icorr) changes from 2.21 μA·cm-2 to 0.55 μA·cm-2 and decreases by 75.11%. Compared with the substrate, the corrosion resistance of laser cladding sample has been improved significantly, and the single pass cladding sample shows the best corrosion behavior.

A total pointing error correction model was built to correct the pointing errors of a level mounting theodolite. Based on the opto-mechanical structure of the level mounting theodolite, a geodetic coordinate system and a pointing coordinate system were built. Then, according to the location relationship between the two coordinate systems, the target coordinate in the geodetic coordinate system was expressed. After a total differential for the target coordinate, an equation was obtained to describe the relation between target coordinate error and longitude & latitude errors of the level mounting theodolite. By considering the three shafting errors to be the major ones of the system, 5 linear coordinate transformations were performed to deduce a total coordinate error equation for three shafting errors.Combined with the total differential equation, a unitive compensation model of pointing error on the three shafting errors was obtained.By linear superposition with the pointing error caused by encoder errors, the total pointing error of the level mounting theodolite was gained. Finally, 46 stars scattered in sky space uniformly were selected to be measured, and the discrete values of the 46 stars’ pointing errors were gained in longitude and latitude dimensions, respectively. After least square fitting, the undetermined coefficients in the total pointing error model were obtained. Experimental results indicate that the total pointing accuracy can rise from 40.1″ to 3.4″ after correction, which meets the accuracy request in a general design.

To obtain a synthetic aperture ladar signal with a high net power rate, a novel method to obtain the Linear Frequency Modulation(LFM) laser signal was presented and the LFM laser experiment was also performed. Detailed Fourier analysis for the driving signal shows that the cavity tuning device can not response the high-order harmonics well. A high-order harmonics response module was introduced into the driving circuit of the cavity resonator to drive the piezoelectric transducer for cavity tuning, so that the trailing edge of the signal was shortened and the net power rate of the signal was improved greatly. Furthermore, the performance of obtained signal was tested and analyzed. The results show that when the FM period is 1 ms, the FM bandwidth is 789 MHz with the FM linearity better than 1.6%. Meanwhile, the net power rate of the signal is raised to 92% from 72%.

For ultraviolet limb imaging spectrometers, a system-level design scheme of the imaging circuit featured by analog-digital separation was presented to avoid the CCD output analog signal transferring between the boards and the interference of digital signals to analog signals.As a result, a 2 LSB noise performance of the CCD signal processing circuit is achieved, which is exactly coincident with the noise specification on the Analog Front End(AFE) datasheet. As the CCD57-10 BI AIMO has no anti-blooming structure and its charges will overflow before getting saturated, a concept about the critical blooming electron counts was defined to replace the saturation electron counts. By raising the high level voltage of CCD transfer clock signals to improve the depth of potential well, the critical blooming electron counts are increased from 30×104 to 60×104, which ensures the highest intensity light signal in the designing range to be correctly detected. To take shorter exposure time to expand dynamic range, the idea of multiple charge-dumping was introduced into the design of drive timing. Without sacrificing the dynamic range of the detector, the minimum exposure time is brought down as small as 19 ms, which is almost an order magnitude lower than that of traditional CCD and ensures the system dynamic range design of 105.

A measurement method of nanoparticle sizes by the variance of temporal coherence of dynamic light scattering was presented to resolve problems of complex algorithms, slow operation, and high costs of traditional dynamic light scattering nano-particle measurement systems and its algorithm and measurement system were studied, respectively. Firstly, the photon correlation spectroscopic theory of dynamic light scattering measurement was introduced, and the concept of coherence degree which includes a temporal coherence factor and a spatial coherence factor was deduced. Then, based on the statistical properties of photoelectric detectors, the variance of the fluctuation of scattering intensity was deduced through analysis of the variance of photon counting. The equations of the variance of intensity fluctuations and the temporal coherence degree were established, and the decay width of a dynamic light scattering signal was obtained by the equations. Finally, the particle diameters of the nano-particles were calculated by the Stokes-Einstein equation.Furthermore, the latex particle sample solution with the diameters of 30, 50, 100 nm and the transmission of 96% was used as the standard sample to be tested in an experiment. The results show that the average values of the measurement error and repeatability error of the new method are 1.84% and 1.76%, respectively, which satisfies the national standard that these errors should be less than 2%, respectively.

To obtain the test data from the experimental research on the compliance factors of the flexure support structure of a large scale reflector used in an airborne image sensor, a compact single-axis flexure hinge experimental setup was proposed and designed. By using this experimental setup, end-bending and pure moment bending experiments of the single-axis flexure hinge could be achieved on a common single-axis Material Testing System(MTS) . This setup utilizes the leverage principle to realize the pure moment loading on the single-axis MTS and takes the side constraint configuration which is composed of several sets of rigid rollers to constrain the pure moment loading sub-assembly.By these ways, the effect of the friction on the load transmission was reduced greatly.Furthermore, the virtual simulation of the pure moment bending process was performed. It indicates that the side force stemmed from the constraint of “L” pushing component is the main influence factor on the bending processing, and the maximum side force can be controlled at 4.783 N. In order to verify the test data, Finite Element Analysis(FEA) and theoretical calculation were employed. The comparison shows that the test results are concident with FEA results and theoretical results well, and test results and theoretical results are slightly smaller than FEA results. Compared with FEA, the maximum relative error of the test data is only 3.87%, which indicates that this experimental setup is a good solution for study of flexure support structures used in optical reflectors.

To design a large aperture rectangular mirror with high stiffness, high strength and good thermal dimensional stability precisely, an analytical mathematical model of the mirror supporting system was established and its modal characteristics were investigated. Firstly, the support location of the mirror was determined according to the conclusion obtained by modal analytical solution that the dynamic and static stiffness is the highest when the three supports keep the mass distribution relatively homogenous. Then, the support structure was analyzed in detail by Finite Element Method(FEM) and a new type of flexible support was presented. By adjusting two parameters, the thinnest thickness and the arc radius of a flexible hinge, the surface figure accuracy of the mirror was modified. Finally, the modal analysis of mirror component was performed. Analytical and experimental results indicate that the surface figure accuracy of the mirror can reach RMS 12.3 nm under the load case of gravity and 4℃ uniform temperature rise, respectively, when the thickness of the flexible hinge is 4 mm and its circular radius is 2 mm. Moreover, its first order natural frequency is 146 Hz, which shows an error under 5% compared to FEA results. Research demonstrates that the mirror structure satisfies all design indexes.

According to the technological bottleneck of elastic style six-axis force sensors, a novel parallel piezoelectric six-axis force/torque sensor was proposed. Firstly, the structure and operating principle of the sensor were presented, and two different arrangements of the eight group piezoelectric quartz crystals for the sensor were put forward. Then their mathematic models were derived, the finite element model of sensor was established, and the output charge sensitivity, coupling interference and the natural frequency of the sensor were analyzed by ANSYS software. Finally, the static and dynamic characteristics of the sensor were calibrated.Research results indicate that the sensor has simple and rational structure, correct mathematic model and simulation way.Moreover, its natural frequency is more than 25 kHz and coupling interference is less than 3% after using the decoupling matrix. These results satisfy the requirements of sensor design targets.

The identification and design of the roll axis servo system with a small journey range in an airborne three-axis gyro body-stabilized platform were discussed and researched by time domain and frequency domain methods. The identification method based on the genetic algorithm was presented by the time domain method and the genetic algorithm was also discussed.Moreover, the opened-loop frequency characteristics of the system were obtained according to the closed-loop frequency characteristic measuring data by the frequency domain method.These methods resolve the difficulty of roll axis servo systems with small journey range in directly measuring opened-loop frequency characteristics. On the basis of the transfer function identified, the control system corrected can meet expected indicators. Experimental results prove that the identification can nicely describe characteristics of the proposed system even in big noises, shorter journey ranges and poor measuring conditions. The closed-loop bandwidth of control system is 121 rad/s and the stability margin is 60° after correcting, which can meet the indicators of an engineer.

As the traditional measuring method based on dielectric coefficients shows cross-sensitivity for multi-parameters in the measurement of oil/water two-phase flows, it can not meet the requirements of real-time optimization control for petroleum production. Therefore, this paper investigates a method to measure multi-parameters with cross-sensitivity by using multi-sensing technology.It presents an inverse model of wavelet neural network with genetic optimization and also researches its identification method. The model overcomes the blindness of initialization weight-value choice in traditional neural networks, provides the abilities of global optimization and nonlinear self-learning, and eliminates the cross-sensitivity of multi-factors. The simulation and experimental results demonstrate the validity and effectiveness of the proposed model and show that the correlation coefficient between the predicted values and calibration values is 0.999 6, which is better than that of BP-NN model. The method has strong generalized capability and robust convergence rate, and can effectively eliminate the influence of the cross-sensitivity of multi-factors and the nonlinearity of sensor self on the measuring precision, and improve the dynamic characteristics and measurement accuracy of sensor systems.

To establish the scientific principles and technical frameworks of the ultra-precision grinding technology, the research work was performed on the processing principles and methods of the ultra-precision gear-grinding to develop the ultra-precision master gears with class 1(ISO 1328-1: 1995). Firstly, according to the optimal forming of an involute and the easy refinement and high structural rigidity of the machine, the gear grinder with flat-faced wheel Y7125 was chosen as the object machine tool. Then the ultra-precision gear-grinding processing was discussed and studied from the aspects of gear blank processing and datum renovation, optimal adjustment of grinder parameters and scientific gear-grinding operation. Finally, the medium module (m2 and m4) ultra-precision master gears with class 1 were developed by a scientific gear-grinding experiment. The research on ultra-precision gear-grinding processing plays a guiding role in the overall improvement of gear manufacturing technology in China.

A solar tracker with intelligent switching ability of working modes is constructed to improve its performance and reliability under several weather conditions.How to choose a proper work mode of the solar tracker is investigated, and the switching method of working mode for the solar tracker is also proposed. The solar tracker has two working modes, open loop mode and closed loop one, and it consists of an imaging unit, a turntable, steppers, the subsystem of micro control unit, a computer and so on. A visual servo scheme is employed for the closed loop mode, while astronomy theory is used for the open loop mode. Moreover, the sun position in the sky is obtained by astronomic algorithms in the case of open loop. The system captures the sky image through an imaging unit, and analyzes weather conditions and external environments through image processing. In experiments, the average tracking error is 0.186 5° for the zenith angle and 0.167 9° for the azimuth angle. The solar tracking experiments indicate that the solar tracker operates reliably, and is able to work under complex weather conditions.

Scanning White-light Interferometry (SWLI) was proposed to measure the complex microscopic surface topography of a dimple fracture of 30CrMnSiA alloy. The Linnik structure was adopted in the system and the 3D image of the dimple fracture was rebuilt by the cosine Fourier analysis algorithm. Experiments show that the longitudinal measuring accuracy is better than 5 nm for a scan range of 120 μm. The cosine Fourier analysis offers strong abilities of phase extraction and noise suppression, and an ideal 3D reconstruction image of the dimple fracture is obtained. The obtained surface data were tested for fractal dimension by a slit island method. At incised height from 40% to 70%, the fractal dimension values range from 1.630 4 to 1.643 2 with an average of 1.641 7 and the standard deviation of 0.012 0, which indicates that the dimple fracture of 30CrMnSiA has a typical fractal character. In conclusion, the SWLI is an effective way to measure the microscopic 3-D fracture surface topography, and it is characterized by high longitudinal precision, long dynamic measuring range and high reconstruction efficiency.

The entire calibration and compensation method of a Miniature Inerial Measurement Unit(MIMU) in high dynamic and overload complicate environments was proposed.Firstly, an error model applied to the complicate application environments was established, which consists of the structure errors, installation misalignment errors and the errors of the MEMS sensors including zero output drift, temperature drift, cross-axis error, nonlinear scale factor error and acceleration effect error of gyroscope. Based on the model, the entire calibration and compensation method was proposed to calibrate 63 error coefficients without calibration of each MEMS inertial sensor separately. Then, the generalized least square algorithm was used to calibrate and calculate the error coefficients. Finally, a MIMU was developed for a flight experiment and was calibrated with this proposed method. Experimental results indicate that the positioning accuracy is improved by 1 order of magnitude. It can satisfy the high dynamic and overload requirements of the MIMU.

Three-dimensional curve fitting and reconstruction of an intelligent structure shape were researched to implement the active monitoring for a high-performance aircraft structure shape. Firstly, the B-spline curve fitting method and recursive algorithm for coordinate fitting were introduced, and algorithm processes and implementation steps as well as numerical simulation analysis were presented. Then, a flexible structure model and its experimental environment were established, and visualization software was developed based on the Visual C++ and OpenGL technology. Furthermore, the curvature detection principle of the orthogonal distributed Fiber Bragg Grating(FBG) sensor array was demonstrated, subsequently, the three-dimensional space shape fitting and reconstruction of the experiment model structure were analyzed and verified. Finally, B-spline fitting algorithm and linear interpolation algorithm were adopted separately to construct a standard circle with the radius of 50 cm. Experimental results show that B-spline fitting algorithm has better graphical reconstruction effects comparing with linear interpolation algorithm, and its standard deviation is 40% that of linear interpolation.As the FBG intelligent flexible structure shows the better shape reconstruction and visualization, it verifies the high precision and low error characteristics of the B-spline algorithm.

For the segmentation of the non-brain tissues and brain tissues from the T2 weighted volumetric Magnetic Resonance Images (MRI), an image extraction method including two levels was proposed to extract brain tissues based on deformable surface models and mathematical morphology. The first level was finished by deformable surface models and region growing according to the brain anatomic, imaging knowledge, and the distribution of MR brain tissues in intensity histogram; the second level was completed by using mathematical morphology to erode the outcome of the first level of brain extraction to obtain more precise results. The experimental results demonstrate that the accuracy rate of the human brain extraction from T2 weighted MRI can reach more than 94%. The algorithm performance evaluation proofs that the proposed method is effective on the extraction of brain tissues from T2 weighted MRIs.

An adaptive double axis motion control system to improve the accuracy of processing platforms was designed by combining the adaptive technology and the interval type-II fuzzy neural network theory.The system controled two field oriented permanent magnet synchronous motors to locate the X-Y double axis motion rotor to track the butterfly contour. Meanwhile, a robust compensator was proposed to confront the Lumped uncertainty, including the inevitable approximation error due to the finite rules of the interval type-II fuzzy neural network, optimal parameter vectors and so on. Finally, the proposed control algorithm was implemented in a TMS320C32 digital signal processor. The experimental results indicate that the butterfly contour tracking performance of the double axis motion control system is improved significantly, and the control system based on the interval type-II fuzzy neural network is more robust than that based on the type-I fuzzy neural network for different uncertainties.

Because of the existence of transition zones in a cold rolling strip surface defect image, gray information and spatial structure information should be combined to segment images to obtain better image results.Therefore, the excess entropy of information entropy and fuzzy set theory were researched.As the excess entropy could be used to measure spatial structure of an image and the characteristic of image gray transition zone could be described well by the fuzzy set, an image threshold segmentation algorithm based on maximal fuzzy excess entropy was proposed.The fuzzy excess entropy was built by the combination of excess entropy and fuzzy set theory and the threshold was determined by the best membership function parameter combination according to the maximal fuzzy excess entropy value. Then, the image was segmented by using the threshold. Finally, the algorithm was compared with Ostu and 1D maximal fuzzy entropy segmentation algorithms. The experiment indicates that the proposed algorithm can extract the defect from a background exactly and can constrain the over-segmentation effectively. The quantificational evaluation of segmented image was performed by the wrong segmentation rate and effective information rate, and the results show that the effective information rate of the algorithm is higher than 82.7%, which is the maximal one among three methods.Meanwhile the wrong segmentation rate is below 2.1%.

For aerial images with poor contrast and color fidelity due to foggy and hazy weathers, this paper proposes a technique of haze removal for aerial degraded images based on the dark-channel prior and the physical model to improve the visibility of vision system in an Unmanned Aerial Vehicle. From the viewpoints of image restoration and image enhancement, the optimized models of global haze removal and self-adapting contract extending are established, respectively. Using the method, a high quality haze-free image can be recovered and the thickness of the haze can be also established. The experimental results on a variety of outdoor haze images demonstrate that it can enhance the contrast and color definition of hazy degraded images fast and efficiently and can achieve satisfactory visual effects. Moreover, the method overcomes the Kaiming He’s drawback of more time consuming, and the aver-age processing time is 10% that of the traditional method.It provides a theoretical reference for the real-time haze removal processing in engineering projects.

The Scale-invariant Feature Transform(SIFT) algorithm was improved in this paper, which could match two pictures and could also compute the object rotation angles in the pictures. Firstly, the SIFT was used to extract two images according to the SIFT feature invariance. Then, the computing process for the main direction of the feature point was analyzed, and the main angle for every key point was recorded. After matching two pictures, the angle difference of main direction for each pair of matched SIFT feature points was calculated and the rotation angle of feature point was obtained. Afterthat, all of the rotation angles of the feature points were analyzed by iterative self-organizing clustering method. Finally, the correct class of samples was selected by the variance and the number of samples within the classe, and the mean of the correct class was used as the final rotation angle of the object. The experiment results indicate that the rotation angle error is within 3° when the image distortion is not significant and it can also well estimate the rotation angle even if the object is partially occluded. Furthermore, in the case of the time complexity does not increase obvionsly, the SIFT can compute the rotation angle of the object, which expands its applications.

In order to reduce the testing risk of an on-board embedded software and improve software developing efficiency, an on-board embedded software test planning method was proposed to plan the whole test process based on a butterfly model. Firstly, the principle of butterfly model was introduced. According to the characteristics of on-board embedded software, the on-board embedded software test planning method was put forward on the basis of the butterfly model test principle, and the advantages of the method were analyzed. Then, by taking an on-board camera embedded software for example, the software development process was planned with this method. Finally, the test results were analyzed. The results indicate that this method could discover and avoid requirement and defect errors about 80 percent and design and defect errors about 60 percent.Therefore, it can reduce software testing risk, shorten software development cycles, and improve the reliability of on-board em-bedded software.

As the origional Fast Hessian which is the most efficient blob feature detection algorithm can not meet requirements of those images in real-time applications to the target recognition, target tracking and so on, an accelerated Fast Hessian multi-scale blob feature detection algorithm is proposed to upgrade the detecting speed of the Fast Hessian. The basic idea of the proposed algorithm is to decrease the number of filter operations and to calculate selectively the values of sample points in the first and last scales for each Octave. Compared to the original one which calculates all the sample point values in these scales, the number of filter operations are distinctly decreased and the consuming time of detecting processing is also reduced. The experiments indicate that the accelerated Fast Hessian algorithm and the original one have the same detection results, but the implementation speed of the accelerated Fast Hessian is upgraded nearly 40% of the original one. It concludes that the accelerated algorithm is much more fit for real-time applications.

Pairwise combinatorial coverage testing which is a significant method for test data generation in functional test fields was discussed. As the method is difficult of solution, a improved method of test data generation for the solution space tree based on intelligent planning was proposed according to the relationship of different factors in the system.The method generates all the available data into a solution space tree, then searches a path from a root node to a leaf node to produce a test data.Experimental results show that this method can test adequately and cover more effect factors of software system by using less test data and can achieve the combinatorial corerage of two factors for test data. As compared with traditional methods, it can obtain the simplest and more complete data sets.In conclusion, it improves the efficiency and veracity of generating test data when this method is taken to a bear space camera software test.

On the basis of imaging characteristics of a reflection stereo vision, an approach for disparity estimation and image restoration of reflection stereo vision was proposed and the double image of the vision system was restored.The reflection stereo vision is a monocular system, and can capture the double image in real time. In experiment, the image was segmented into the rectangle observing blocks firstly, then, the cepstrum analysis was used to estimate the displacement amount (disparity) of each pixel in the double image by detecting the two symmetrical peaks from the cepstrum domain and to calculate the 3D space distance of a target point. Finally, the disparity was used to combine an inverse filter to restore the double image and to obtain a good scene image. Experimental results show that the method is effective for disparity estimation and image restoration, and the accuracy of the disparity estimation has improved by 15%-20%.