A dual-channel photonic crystal filter model (AB)m(LC)H(LC)(BA)m was proposed with nematic liquid crystal as defect layer. Modulation characteristics of external electric field to the filter were studied by method of combining transmission matrix with electro-optic effect of nematic liquid crystal. The results show that the linear relationship between the central frequency of the double filter channel shows that the modulation effect of the voltage on the filter frequency is very good. The response sensitivity of the two channel filter bandwidth to the external field voltage is different, and the bandwidth of long wave channel is more sensitive to the voltage than the other one. When the voltage is small enough, the modulation range of the voltage to the filter bandwidth is large, and the modulation law is obvious.The filter bandwidth can be adjusted by selecting the appropriate voltage range. These characteristics provide definite reference to design of new optical filter.

Photon detection has the advantages of high sensitivity, low system power consumption and long detection distance. And its performance is easily affected by the atmospheric environment. A photon detection model affected by the turbulence was established, and the modulated Poisson distribution of the echo photon was defined. The modulated Poisson distribution degenerated into Poisson distribution when the number of echo photons was less than 0.1. The range walk error model of photon detection was established to reconstruct the echo and calculate the range walk error of photon detection. The characteristics of photon detection echo and range walk error affected by turbulence were discussed. The results show that under the influence of turbulence, the first photon effect of photon detection is weakened; when the turbulence is 10-16 m-2/3 and the number of echo photons is 2, the range walk error of photon detection is reduced by 0.41 cm and the range precision decreased by 0.23 cm, compared with when there is no atmospheric turbulence. When the turbulence is 5×10-15 m-2/3, the detection probability of photon detection is reduced by 78.18%on average, the range walk error decreases by an average of 91.49%, and the range precision drops by a maximum of 85.77%.

Aerosol is one of the important contribution sources of air pollution, which affects the global radiation balance and climate change. At present, active and passive remote sensing satellites(such as MODIS and CALIPSO) have relative low temporal resolution to obtain global aerosol information.MERRA-2, a global reanalysis data, can obtain high temporal resolution 3-D aerosol distribution, but the quality of MERRA -2 aerosol product is still need to be evaluated. Thus, this study verified the accuracy of MERRA-2 data in central and eastern China(especially in areas with heavy human pollution) by cross-comparing the three-dimensional profiles of extinction coefficient of MERRA-2 and CALIPSO. Results show that the extinction coefficients of MERRA-2 and CALIPSO have a good overall correlation(0.79), of which the two datasets have the highest correlation in autumn(0.82), followed by summer and winter(0.81 and 0.80), and the lowest correlation in spring(0.77). As the atmosphere becomes cleaner along with the altitude increasing, the correlation coefficient between MERRA-2 and CALIPSO declined from 0.78 to 0.52, and the standard deviation declined from 0.028 to 0.0014. The above results demonstrate that MERRA-2 can provide more accurate 3-D aerosol information under heavy pollution. Meanwhile, the extinction coefficients of MERRA-2 and CALIPSO are not greatly affected by the type of aerosol.

Rayleigh Doppler Lidar was often used for horizontal wind speed and temperature observation in stratosphere and middle atmosphere. University of Science and Technology of China introduced a Doppler frequency calibration device in the newly developed Lidar system, which can achieve high-precision observation of the horizontal wind speed and temperature of 15-70 km of the atmosphere under clear sky. The continuous observation data of the horizontal wind speed and temperature of the Lidar system was analyzed in the northwest area of China from September to October, 2018. The gravity wave parameters, such as the vertical wavelength, natural period, horizontal and vertical propagation directions, were reached by extracting wind speed and temperature disturbance, spectrum analysis, gravity wave model parameter fitting and Hodograph. A total of 384 gravity waves were extracted from 224 h of observation data. In these waves, 53.9% of the vertical propagation direction is upward. Fluctuation amplitude of wind speed and temperature is concentrated at 10-20 m/s and 10-20 K. More than 70% of the vertical wavelength is in a range of 5-10 km, and fluctuations of more than 15 km rarely occurred. The horizontal wavelength is centered at 100 km to 200 km, and the dominant intrinsic period is less than 5 h. The horizontal propagation direction is dominated by the opposing directions of 20 ° west-northwest and 20° east-south. The distribution characteristics of gravity waves are consistent with the local dominant wind direction and the distribution of surrounding mountains.

Micro Pulse Lidar (MPL) has been widely used in the measurement of the vertical distribution of aerosols, while it is usually set to a fixed station and cannot observe the whole spatial distribution. In this study, the MiniMPL integrated in the car was used to carry out cruise observations of Shijiazhuang and its surrounding areas, and the continuous three-dimensional spatial distribution of aerosols in the area was mapped out. Results revealed that: (1) the air pollution in this area is influenced by the topography, the west Taihang Mountains block the transport of aerosols in the Fen River Basin, aerosols converge in the lowlands in front of east Taihang Mountains and spread eastward to the vast plains; (2) the aerosols convergence zone is formed from the northeast to the south of Shijiazhuang, the concentration of air pollutants gradually increases from the northeast to the south, in addition, there is a clear boundary between the polluted zone and the clean zone that in the northwest part of the Taihang Mountains and Yanshan Mountains; (3) the strong vertical dispersion of aerosols and high irregularity of particulate matter in the outer space of Shijiazhuang may result from multi-source mixing, and the differences of the aerosols height in the four directions may be caused by the uneven distribution of pollution sources.

From November 1, 2017 to May 15, 2018, the micro-pulse lidar level detection was carried out at the Quancheng Square substation in Jinan. During the experiment, the visibility meter, conventional pollutant monitoring and PM2.5 chemical composition monitoring were simultaneously observed. In this experiment, the visibility meter was defined as the reference layer. And the horizontal aerosol extinction coefficient was calculated by using the Fernald method. The relationship among PM2.5 concentration, extinction coefficient and PM2.5 chemical components was established by linear regression model. The inversion accuracy of the relationship was studied by trend comparison, correlation coefficient and standardized mean deviation. The results show that the extinction coefficient is more accurate by combining with visibility meter; the inverse extinction coefficient has a good positive correlation with PM2.5 mass concentration (correlation coefficient R=0.75), and the extinction coefficient and PM2.5 mass concentration are all related to O■,NH■■ , OC, NO■■, EC in PM2.5 chemical components. The relationship among PM2.5 concentration, extinction coefficient and chemical components of PM2.5 is established by linear regression model. The mass concentration of PM2.5 retrieved from the quantitative relationship is basically consistent with the actual monitoring value, and the correlation number is 0.83. The average standard deviation is 18%, and the relative deviation is small.

As one of the important means of active remote sensing of atmosphere, lidar is widely used in aerosol and cloud detection. Based on the simple multiscale algorithm for layer detection, a cloud detection algorithm was proposed that could improve the accuracy of the cloud-base height when the SNR was greater than 5 by adding the threshold of the number of scales. By simulating the echo signal of multi-peak inside the cloud, the algorithm obtained the threshold range of the optimization scale. At the same time, by adding feature segment merge, the disadvantage of the simple multiscale method, missing the single-layer with multipeak, could be improved. The effective 532 nm Mie lidar data was processed which had more than 75 min detection time and stable structure from May to July 2019 by using differential zero-crossing method, simple multiscale method and improved algorithm were proposed respectively, and based on the cloud-base height of differential zero-crossing method, the average error of the mean square root obtained by the improved algorithm decreased by 32.65%, the uncertainty average decreased by 33.80%. Then the effectiveness of the improved algorithm in improving the accuracy of the cloud-base height is proved.

In order to investigate the effective distance and retrieval method of CCD laser beam imaging technique in detecting ozone concentration profile, a hardware detection system was designed. This system employed the 266 nm and 280 nm continuous-wave laser as differential absorption lidar transmitter. The signal-to-noise ratio (SNR) of detection echo signal was simulated based on the main parameters, such as laser power, exposure time, the effective area of CCD camera, aerosol concentration and ozone concentration. Supposed that aerosol is clean, pollutant and heavy pollutant types and ozone concentration is low and high cases, with the uncertainty of less than 10%, theoretical analysis results show that the effective detection distance is more than 2 km under the condition of three different aerosols cases with low ozone concentration; the effective detection distance is still more than 2 km under the clean and pollutant aerosol cases and more than 1.7 km under heavy pollutant aerosol with high ozone concentration; the retrieval ozone concentration profiles are consistent to the assumed ones. The simulation results are helpful to understand the optimal parameters of CCD laser beam imaging technique in detecting ozone concentration profile and then realize effective detection to ozone concentration profile.

Doppler wind lidar is an effective remote wind measurement instrument with high temporal and spatial resolution. However, due to the interference signal reflected by raindrops, it is a challenge to perform precise wind profile measurements under rainy conditions, but it also provides a possibility for detecting rainfall. In this work, a compact all-fiber coherent Doppler lidar(CDL) operating at wavelength of 1.5 μm was applied for simultaneous wind and rainfall precipitation detection. Due to the ability of precise spectrum measurement, both aerosol and rainfall signals can be detected by the CDL under rainy conditions. The echo signals from aerosols and raindrops with different speeds will cause two peaks in the Doppler spectrum, so that the spectrum width can be used to identified rainfall events. A two-component Gaussian model was applied to fit the spectrum and two velocities were obtained. The comparison with the results of the micro rain radar verifies the CDL′s ability of rain measurement, meanwhile, the false detection probability of wind speed in the rainy conditions is also reduced.

A scalable visual measurement was presented to meet the technical requirements of target location under large size and long distance non-contact conditions. Many visual measurement methods seek to obtain higher accuracy on the improvement of PnP algorithm, but do not consider the impact of increasing the number of feature points. Thus, a cooperative target design by quantitatively expanding the number of feature points to achieve higher positioning accuracy was proposed. Firstly, through the establishment of measurement model and accuracy analysis, the function analysis formula of measurement accuracy of n-point target was obtained by using error propagation theory. Then the accuracy of the model was verified through simulation experiments, and the effect of the arrangement of the feature points of the cooperative target on accuracy was further discussed. Subsequently, the target layout and structural parameters were optimized. The experimental results show that the reprojection error of the designed 16-point ring array target is reduced by about 50.3% compared with the same size P4P target, and the stereo target further reduces the reprojection error by 39.2%. This technology can be used not only for target positioning in special environments, but also for unit state monitoring in complex systems.

A fiber grating (FBG) vibration sensor based on a flexible hinge suitable for low-frequency vibration measurement was proposed, which realized high-precision measurement of low-frequency signals within a certain range in response to the needs of low-frequency vibration detection. According to the ideal structure model of the sensor, combined with the torque balance equation and the Lagrangian equation, the calculation method of the sensitivity and natural frequency of the sensor was given, and the structure of the sensor was analyzed by finite element using ANSYS software, and the first-order resonance frequency was obtained as 92.457 Hz. Through the performance test, the sensor has good flatness in the frequency range of 0.5 to 55 Hz, and the natural frequency is 91 Hz, which is consistent with the finite element analysis result. The sensitivity is 1.94 nm/g and the linearity is 99.99%. The longitudinal sensitivity is much better than the lateral sensitivity, and it has strong lateral anti-interference ability.

In order to study the propulsion performance of metals ablated by nanosecond laser with 532 nm and 1 064 nm wavelength, 8 ns and 15 ns fundamental frequency, the mass removal, specific impulse, momentum coupling coefficient and laser energy coefficient of metals were measured under background pressure of 40 Pa. The results show that the mass removal of iron is largest, while yttrium has the lowest mass removal. The mass removal of metals during laser ablation is determined by laser absorption, melting point and boiling point of metals. Under conditions of larger power density(＞2.71×1010 W/cm2), the specific impulse and laser energy coefficient of yttrium are largest, while the iron has the lowest specific impulse and laser energy coefficient. The changes of metal ablation characteristics are explained by mechanisms of ground state atoms excited to form plasma and irradiation laser shielded by plasma. The research is helpful to selection of metals using for laser ablation.

The echo signals of airborne LiDAR based on streak-tube by neural network were classified. The typical echo signals of four different targets were compared and analyzed, and the intensity and morphology in echo signal were extracted as features. A BP neural network classifier were constructed by MATLAB. The effect of the number of hidden layer neurons, neural network training algorithms and training sample number on the performances of the classifier was compared and selected. The test results of echo signals using this classifier demonstrate that the accuracy of this classifier can reach 96% and Kappa coefficient is 0.95, which is capable to classify the echo signals accurately.

In order to meet the application needs of unmanned driving and vehicle obstacle avoidance, imaging lidar usually needs to identify target signals in a large dynamic range of several meters to several hundred meters. Limited by volume and cost, lidar usually does not have automatic gain control function. Therefore, high precision measurement of saturated to unsaturated echo signals is one of the problems that imaging lidar needs to solve. Traditional imaging lidars mostly use the TDC method. This method is effective for the saturated waveforms of the targets nearby, but has limited imaging ability for unsaturated echo signals far away. Therefore, a method of full waveform matching based on high speed ADC was proposed. The high-precision peak position of the echo pulse was obtained by sliding matching the echo pulse and the template waveform. And an optimized fast algorithm on FPGA was implemented, which improved the execution speed. The accuracy of ranging was ensured and the effective imaging range of the imaging lidar was increased. Through a large number of experiments, the imaging lidar has a ranging accuracy of 2.7 cm and an effective imaging range of more than 200 meters.

In the underwater application of lidar, the underwater transmission of laser light is complicated and variable due to the absorption and scattering of light by water, especially the noise caused by backscattering will reduce the contrast of the target, and the target echo even will be drown, resulting in big challenges of water downward laser echo detection. The subcarrier modulation technology can effectively improve the underwater detection performance of the lidar system and increase the signal-to-noise ratio of the target echo signal. Based on the Monte Carlo method, an improved underwater transmission model of a subcarrier-modulated pulsed lidar was proposed. The effects of seawater attenuation coefficient and passband on system performance were studied. Simulation results show that compared with traditional lidar, the environmental parameters of the water body and the selection of the passband during filtering have a greater impact on the underwater detection performance of the system; the subcarrier modulation technology can improve the target signal-to-noise ratio and measurement accuracy, it effectively improve the system performance.

The inverse synthetic aperture ladar imaging technique was applied into laser wavelengths and airborne platform, and realized the sparse reconstruction on conical target under micro-motions, with range and azimuth resolution of an order of centimeter. By establishing the imaging model of airborne inverse synthetic aperture ladar, the formations of target′s range distance and ladar echo were analyzed, and the feasibility of detection on conical target by inverse synthetic aperture ladar and airborne platform was verified. Aiming at the problems of high repetition frequency, complex system and data redundancy, a sparse imaging algorithm was proposed based on spatial compressive sensing, to form a linear observation model under the condition of sparse sampling, and transform the imaging problem into an optimization solution of sparse signal on l1 norm organization, in order to realize the reconstruction of target with micro-motions. The experiments of electromagnetic calculation demonstrate the validity of proposed method.

The complex water environment, especially the absorption and scattering of particles in the water, causes a large number of backscattered clutter at the receiver of the underwater lidar ranging system. As the distance increases, the target echo will gradually drown in clutter. The modulation technology can effectively improve the contrast of the target and achieve the range extension of underwater lidar. A Monte Carlo simulation model of photon transmission in seawater was firstly established, and then the modulation technology which combining the advantages of lidar and microwave radar was studied. Then the detailed steps of transmitting, receiving and related processing was given when a pseudo-random sequence was used as a sub-carrier. Finally, four different types of sub-carriers(square wave, sawtooth wave, sine wave, and pseudo-random sequence wave) were used for simulation studies with the different water qualities and emission pulse widths. The results show that the pseudo-random sequence has the best performance in improving the contrast of the target.

ICESat-2 is the first photon-counting lidar altimeter satellite in the world. Its ATLAS can obtain high-precision laser foot points covering the world, which helps to improve the positional accuracy of stereo satellite images without ground control points (GCPs). A joint block adjustment method for spaceborne laser altimetry data and stereo satellite images was designed and implemented. First, a multi-condition filter was proposed to select high-precision ATLAS laser altimetry points, and then the selected ATLAS laser altimetry points were used as control points with error for joint adjustment calculations,so that the satellite images orientation parameters were optimized. Experimental results on the ZY3-02 satellite images and ATLAS ATL03 data of Zhengzhou prove that the proposed method can respectively improve the vertical accuracy and the plane accuracy of stereo satellite images by 60.6% and 56.4%. The experiments discussed above demonstrate that the proposed method can effectively improve the positional accuracy of stereo satellite image without GCPs.

The setting of the jamming lead has an important influence on the success probability of laser angle deception jamming. By establishing the probability model of the guidance signal in the laser angle deception jamming and the probability model of the jamming signal being accepted by the wave gate, the optimal jamming lead formula was deduced and verified. The influence of different factors on the optimal jamming lead setting was studied by simulation. The simulation results show that the gate width, the guidance signal interval error and the jamming signal interval error have great influence on the setting of the optimal jamming lead. The optimal jamming lead is proportional to the width of the gate. The larger the guidance signal pulse interval error is, the smaller the jamming signal pulse interval error is, and the larger the optimal jamming lead is. The jamming probability of laser angle deception jamming can be greatly improved by setting the appropriate jamming lead.

In view of the low accuracy of the estimation of forest biomass by spaceborne lidar waveform data, Ice, Cloud, and land Elevation Satellite (ICESat)/Geoscience Laser Altimeter System, GLAS) echo waveform was used as an example, by modifying the background noise threshold of the GLAS echo waveform, identifying the GLAS canopy echo waveform, setting the position of the start and end points of the canopy echo waveform, the canopy echo energy parameter Eca, and based on the Laser Profile Array(LPA) data was corrected to Eca to obtain the corrected canopy echo waveform energy value parameter ECca, and then the aboveground biomass(AGB) was estimated. The research results show that the energy accuracy parameter Eca of the GLAS canopy echo waveform before and after correction by LPA data was 0.87 and 0.89, respectively, and the RMSE was 17.28 t/ha and 15.76 t/ha, respectively. The research results show that the canopy echo energy value parameter proposed in this paper can effectively estimate forest biomass, and the canopy echo waveform energy value parameter Eca can be improved by LPA data to improve the estimation accuracy of forest biomass.

A high energy 1 064 nm Q-switched laser was obtained by end pumping Nd:YAG crystal with a vertical array of semiconductor laser diodes (LD). According to the physical characteristics of linear change of LD wavelength with temperature, the wavelength of pump light deviated from absorption peak of Nd:YAG by adjusting the temperature of LD, and the gain of pumping end of laser medium was reduced, which effectively suppressed the self-excited oscillation restricting the output energy of Q-switched laser. 1 064 nm laser with a maximum output of 95 mJ was generated under pump energy of 530 mJ. The corresponding optical conversion efficiency was 18% and the Q-switching ratio of dynamic to static was 73%. The laser pulse width was 8 ns and divergence angle was 2.5 mrad.

Airborne lidar bathymetry(ALB) was an effective technology for simultaneous terrain and shallow water bathymetry mapping in coastal zones. Up to present, there had not been mature and practical ALB systems in China, and the experience in large-scale airborne lidar bathymetry production and application was also very little. It was urgent to carry out airborne lidar bathymetry system application test. Using the CZMIL Nova II system, we carried out the application test of the ALB in terms of maximum depth penetration, depth measurement accuracy, and measurement efficiency in Hainan, Guangdong, Guangxi which were located in the northern part of the South China Sea. The results show that the maximum depth penetration of the three typical areas of Wuzhizhou Island in Hainan, Weizhou Island in Guangxi, and Chixi Town in Taishan City in Guangdong reached 30, 16, and 3 m respectively, accorded with the system′s nominal maximum depth penetration capacity. The maximum depth penetration mainly depends on water clarity, bottom reflectivity, flight altitude and other environmental factors. The depth measurement accuracy in Wuzhizhou Island, Hainan is 0.369 m at the 30 m depth, accorded with the system′s nominal accuracy. Based on the Yun-12E aircraft and conventional flight parameters, the effective coverage area is about 100 km2 and the flight line length is about 500 km in a flight mission, the annual coverage area is about 8 000-12 000 km2 and the annual flight line length is about 40 000~60 000 km. This was the large-scale application test of the airborne lidar bathymetry system in the coastal zone of China and it would have important reference significance and scientific value for the research and application of domestic airborne lidar bathymetry technology.

Metastable helium Lidar detects the metastable helium density in the thermosphere and exosphere by emitting 1 083.0 nm laser, which is locked at the resonance lines of metastable helium atoms. The detecting range can cover the altitude from 200 km to 1 000 km. In this height range, metastable helium density was quite low, usually lower than 1 cm-3. University of Science and Technology of China (USTC) was now developing metastable helium resonance fluorescence lidar based on six 1 m-diameter telescopes array and high power 1 083 nm pulse laser. The effective area of this telescope array was 4.8 m2; The high power pulsed 1 083 nm laser employed optical parametric oscillation (OPO) and optical parametric amplification (OPA) techniques, used 532 nm pulses as pumping laser, and emited single mode 1 083 nm pulses with a frequency bandwidth of300 MHz; The receiving system employed a InGaAs single-photon-counting detector with deep cooling (-120 ℃), the quantum efficiency was higher than 30% accompanied with a dark count lower than 100 s-1; The background noise was filtered by series-connecting of interference filter and FPI ultra-narrow-band filter. The final filtering bandwidth was 10 GHz and the peak transmittance was 55%. Based on the above key performance of the system, the SNR would be about 400 in the strong-signal season and 10 in the weak-signal season.

Aerosols have been identified as one of the uncertainties in the evaluation of radiative forcing. Lidar is effective remote sensing tool for aerosols observation. The vehicle-mounted lidar was used to continuously observe the particle concentration in Anping County by combining fixed vertical detection and mobile navigation monitoring. The results of fixed detection are in line with the trend of air quality rising first and then falling in accordance with the trend of the Ministry of Environmental Protection. Mobile navigation detection can quickly and accurately locate the location of the pollution source. Finally, the results of navigation monitoring before and after spraying the dust suppressant are compared, and it is concluded that spraying the dust suppressant can effectively reduce the concentration of particulate matter.

The space-time registration technology between the measurement sensors of the mobile LiDAR measurement system is the core technology of the entire system design, and the problem of time synchronization is particularly critical. Aiming at the problem of accurate time synchronization between the inertial navigation system and the LiDAR and panoramic camera, combined with the Beidou navigation grant time system, a method was proposed which designed a time synchronization device to provide space-time registration information for vehicle-mounted measurement platforms that integrate LiDAR, inertial navigation system, panoramic camera and other mapping sensors. The vehicle-based mobile measurement system with full independent intellectual property rights was taken as the research object. Through this measurement system the LiDAR point cloud data was collected. Using the designed time synchronization device, the prepared upper software were combined with Beidou satellite navigation system for accurate time synchronization. Finally, the accuracy of time synchronization reaches 0.1 ?滋s, which verifies the feasibility of the design method of time synchronization device.

Multispectral full-waveform lidar can simultaneously acquire full-waveform comprehensive information including implicit spectrum information. Waveform decomposition is a key technology that affects the accuracy of multiple echo parameter inversion. A waveform decomposition method suitable for multispectral lidar data was proposed, which flexibly selected fitting sub-functions for each channel and made full use of the spatial correlation between multiple channels to achieve accurate extraction of the spatial position and spectral information of the target; the mapping relationship between the spectral space and the RGB color space was established to achieve the target three-dimensional color reconstruction. Experimental test results show that this method can achieve full-time 3D color reconstruction of the target, effectively overcome the influence of weak echo signal loss and ambient light interference, and effectively improve the accuracy of target ranging and color information inversion. The experimental results show that the method effectively overcomes the interference of ambient light and can achieve the target three-dimensional color reconstruction throughout the day. By comparing with existing methods, the two error evaluation indexes of error standard deviation and determination coefficient R2 both verify that the proposed method has higher ranging accuracy and color information inversion accuracy. In addition, the experimental results verify that the optimal pulse accumulation number can effectively balance the data processing accuracy and scanning efficiency. In the future, the proposed method can be extended to other lidar systems to further improve the accuracy of target 3D color reconstruction by optimizing the laser detection bands.

Lidar systems based on single-photon avalanched diode(SPAD) array detector not only have high sensitivity and time resolution, but also have high detection rate. It has wide application prospects in the fields of remote sensing mapping, target recognition, spacecraft landing and so on. Due to the shortcomings of the current SPAD array devices, such as the small pixel format and the hot pixel problems, the quality of the three-dimensional(3D) images obtained by directly using the SPAD array is poor and the imaging area is small. A continuous scanning three-dimensional imaging method with SPAD array based on pixel multiplexing was proposed to expand the imaging area. At the same time, the multiplexing of pixel information could increase the cumulative detection times of the same target point, thereby improving the depth measurement accuracy and the 3D imaging quality. A push-broom imaging experimental system was established based on a 32×32 SPAD array. By performing a one-dimensional continuous scanning imaging experiment on a target 3.3 m away, a fast and large-scale photon counting 3D imaging(32 rows×520 columns/7 s) was achieved. Meanwhile, the influence of hot pixels on the imaging results is effectively alleviated, and the plane accuracy of the three-dimensional image achieves 2.3 mm.

Femtosecond optical frequency combs have the characteristics of ultra-wide spectral range, ultra-narrow pulse width and ultra-stable pulse sequence interval. Not only can the periodic pulses in the time domain be used to achieve a large-scale time-of-flight ranging, but also the coherent spectrum in the frequency domain can obtain high-precision interference information. In the field of large-scale and high-precision distance measurement, the optical frequency comb can make full use of its advantages. The target distance measured by the system is closely related to the system parameters and atmospheric environmental factors. The atmospheric environmental factors affect the measured optical path through the refractive index. It is necessary to simulate the influence of the above factors on the ranging precision. The dual-optical comb method was selected as the high-precision ranging technology solution. The effect of system parameters and atmospheric environmental factors on the ranging precision was deduced in theory and simulated in detail, The simulation results show that as the repetition frequency increases, the effect of the jitter of the repetition frequency on the ranging result will decrease; reducing the repetition frequency difference between the two optical combs can improve the ranging precision; the influence of atmospheric environmental temperature jitter on the measurement distance is 1 μm/(m·℃); The influence of the atmospheric pressure jitter on the measurement distance is 2.6 μm/(m·kPa); the influence of atmospheric humidity on distance measurement is small. The simulation results provide theoretical guidance for the high-precision optical frequency comb distance measurement experimental system, which can provide theoretical guidance for the high-precision optical frequency comb distance measurement experimental system.

Compressed photon-counting radar is a combination of photon-counting radar technology and single-pixel imaging technology. It has the advantages of low cost and ultra-high sensitivity. However, the reconstruction of high-resolution imaging cost a large number of measurements and iterative calculations, resulting in a long imaging time. At present, deep learning compression reconstruction network has been proved to be able to avoid iterative computation and achieve rapid compression measurement reconstruction, But the deep learning compression reconstruction network reported in the literature uses traditional image processing database of noise-free pictures or adding Gaussian noise to the pictures to train the network, the network is applied to the actual compressed photon counting radar system, the performance needs to be further verified. A synchronous control measurement module based on FPGA was independently designed, a compressed photon-counting radar system was built, a Monte Carlo simulation of compressed photon-counting radar system to produce training data was proposed, and a deep learning compression reconstruction network DFC-Net was designed for joint optimization of sampling and reconstruction. The experimental results show that: at 10%, 15%, 20%, 25%, and 30% sampling rate, the reconstruction performance of DFC-Net is better than the existing reconstruction network Dr2-Net and the traditional compression reconstruction algorithm TVAL3.

The study of the vertical distribution of marine matters is essential for maintaining the health of marine ecosystems and utilization of marine living resources. Compared to other technologies the great advantage of ocean Lidar is that it efficiently provide all-weather perception capability of the vertical distribution of marine matters over a large area. The application status of Lidar in detection of marine materials at home and abroad, including seawater turbidity, chlorophyll concentration, CDOM (Chromophoric dissolved organic matter), and the distribution of fish and zooplankton communities was analyzed. The problems existing in these applications were discussed by considering the detection mechanisms. Beside, this review provided some outlooks on the development direction of ocean Lidar for obtaining the vertical distribution of marine matters.

Full waveform data is one of the core data of GF-7 satellite laser altimeter. In this paper, based on the multi-track measurement data of the GF-7 satellite laser altimeter, the quality of waveform data was evaluated by quality evaluation parameters such as waveform background noise, kurtosis and skewness, signal-to-noise ratio, Gaussian characteristic parameters and other quality evaluation parameters, and the potential of terrain feature extraction from waveform data was analyzed by using Gaussian decomposition of building height calculation and inversion of surface slope. The experimental results show that the overall availability of echo data in the test area is 72.59%, the quality of transmission waveform and echo waveform data is good and stable, and the signal-to-noise ratio is slightly better than ICESat/GLAS. In view of the phenomenon of "flat top" and "undershoot" in waveform data, a preliminary processing method is given, and the selection of waveform data is optimized by combining high and low gain. The waveform data can well reflect the terrain characteristics of ground objects, and the echo waveforms of different ground objects present corresponding waveform characteristics. The accuracy of buildings height deduced from waveform data can reach decimeter level.The related conclusions have reference value for the application of laser altimeter data of GF-7 satellite.

The target detection probability and target reduction degree of the three-dimensional reconstruction algorithm of Geiger mode Avalanche Photo Diode(Gm-APD) laser imaging radar are greatly affected by the preprocessing. Improper preprocessing will cause the target to be missing and reduce the signal-to-noise ratio. Preprocessing usually uses Gaussian function or double Gaussian function for convolution, and improper variance in the filter function will also lead to lower target restoration. To improve this problem, the influence of the double Gaussian function preprocessing variance combination on the detection probability was analyzed. Then, Monte Carlo simulation was used to analyze the best variance combination varies in different background noise under fixed laser energy and frame number. The best variance model was obtained by fitting. Finally,reality experiment was used for verification, and the best variance combination obtained by the variance model can reach more than 90% of the best target reduction degree of actual reality. This model has theoretical guidance for actual signal processing.

The three-dimensional pose information of target is more and more widely used in the fields of target motion analysis, target recognition and target tracking. The K-means algorithm based on distance was used to classify the point normal vector by the existing OPDVA algorithm, then the positive direction of the target coordinate system MCS was determined, and the three-dimensional pose angle of the target was obtained. In view of the unsatisfactory effect of point normal vector classification, a three-dimensional pose estimation algorithm based on mean shift point normal vector classification (PEMSPNC) was proposed. The Mean Shift algorithm was used, which did not depend on the initial parameter setting and based on density clustering, to classify the normal vectors of different plane points with different density distribution, and to find the normal vector of the point with the maximum density as the representative normal vector of each class to determine the positive direction of MCS, then the pose angle of the target was calculated, and the target size according to target pose estimation results was computered. The rectangle fitting method, OPDVA and PEMSPNC algorithm were used to test the simulated and measured range profiles. The results show that the pose estimation error obtained by using PEMSPNC algorithm is the smallest, and compared with OPDVA algorithm, the average error is reduced by 0.4434 °, and has a good processing result for the measured data.

In recent years, Nonnegative Matrix Factorization (NMF) methods for hyperspectral image unmixing have attracted widespread attention. However, due to the non-convexity of NMF problem, it cannot guarantee the uniqueness of the solution, and it is easy to fall into local minima. In order to reduce the solution space of NMF problem and improve the unmixing accuracy, a new method of reweighted sparse NMF (ARSNMF) was proposed. Firstly, considering the sparsity of abundance matrix, the sparse constraint was added to the NMF model. Then, considering that the calculation of the problem was complex and not easy to be optimized, it was converted into a form of reweighted sparse constraint, which not only achieved the sparse effect, but also solved the problem that was difficult to solve. In order to improve the convergence speed of the algorithm, the Alternating Direction Method of Multipliers (ADMM) was used to optimize the model, and the objective function was divided into several sub-problems for independent solution. Experiments based on simulation data and real data verify the effectiveness of the proposed algorithm.

According to the application requirements of target recognition, task programming and template preparation, an algorithm of scene identifiability analysis based on evidential networks was proposed to realize the quantitative analysis of the scene identifiability degree. After having acquired the support data of the research area and setting the imaging parameters, a certain number of salient ground objects were extracted as the scene nodes from the data. Then, the identifiability degree of each extracted object was assessed from three aspects, including scale significance, shape uniqueness and visualization. After that, the conditional belief function which represented the mutual support degree between scene nodes was defined by the contour point number of the extracted objects. Finally, the analysis results of the scene identifiability were obtained by the reasoning and fusion ability of evidential networks. Experimental results demonstrate that the algorithm of scene identifiability analysis is reasonable and effective, which meets the requirements of mission planning and thus exhibits great practical value.

The automatic inspection system of offshore booster station is an intelligent inspection robot system for the whole offshore booster station. It mainly completes the automatic inspection of cables, equipment, instruments and other devices in offshore booster station. The color of the indicator light is often used to judge the running state of the equipment, which is a key task of automatic inspection. Due to the semi-transparency of the indicator light, it has the problem of reflecting light under uneven illumination, which makes the color and brightness information and its non-uniform characteristics mixed, resulting in the "data insensitive characteristics" of the indicator light image. The uniformity of the traditional RGB color space is poor, so the color detection of the indicator light relying only on the RGB color space is often false. Compared with RGB space, HSI space can express the light and shade, tone and brightness of colors intuitively, making it convenient to compare colors. Therefore, a color detection method that converted the indicator image from the traditional RGB color space to HSI space was proposed and the H component was extracted to calculate the histogram similarity. The experimental results show that the proposed method greatly improves the color detection accuracy of the indicator light, and has strong anti-interference ability especially under uneven illumination.

In the process of "spectral reconstruction" of quantum dot spectroscopy, the basis function is needed to construct regression model. The traditional reconstruction methods can not capture the high frequency information of the spectrum, and the inversion accuracy is low. To overcome the shortcomings of traditional methods, a special structure of basis function was proposed, to facilitate the fitting of low-frequency macro-contour information and high-frequency micro-detail information of the spectrum to be measured. At the same time, the adjustment mechanism of the parameters of the basis function will also be studied, so that the parameters of the basis function can be adjusted adaptively. Individualized basis functions were customized for each spectrum to maximize the frequency information of the function to be measured. The test results show that the inversion error of this method is about 50%-70% and 90% of the error of the traditional method for narrow-band and broadband spectra, respectively, and it has a certain anti-noise ability, which lays a foundation for the practical application of quantum dot spectroscopy technology.

In order to solve the problem of insulator missing detection and inaccurate positioning caused by small proportion of insulators and complex background in infrared power image, a novel insulator detection network: Feature Selection YOLOv3(FS-YOLOv3) was proposed. The proposed FS-YOLOv3 added pyramid feature attention network to the top-down sampling process of the original pyramid shaped YOLOv3 network. The pyramid feature attention network calculated the feature weight matrix based on the high-level semantic feature map of YOLOv3, and used the feature weight matrix to filter out the redundancy of low-level detail features of the network. Finally, the low-level feature map and the high-level semantic feature map after feature filtering were connected in series to obtain the feature map with both accurate insulator detail information and rich high-level semantic information. The experimental results show that the detection accuracy of the proposed method is better than that of the original YOLOv3 network, and retains the good real-time characteristics of the original network.

With the continuous promotion of object detection technology in electric power inspection tasks, automatic analysis of various images collected by electric power inspection has become one of the current hot research directions in electric power enterprises. Traditional target detection methods are mostly based on machine learning, and the detection accuracy in complex scenes needs to be further improved. The image-based deep learning method is widely used in power inspection target detection due to its ideal detection accuracy and environmental adaptability. Aiming at the problems of poor image quality, complex background, poor contrast, etc. collected in complex scenes of electric power inspection, an infrared image object detection method based on regional convolutional neural network fused with histogram of image orientation gradients(HOG-RCNN) was proposed. Before the image enters the RCNN, HOG feature extraction was performed on the input image for helping RCNN to select candidate regions. Algorithm experiments show that the detection effect of the method proposed is better than that of a separate RCNN network.

Pedestrian detection is a research hotspot and a difficult issue in the computer vision such as the Intelligent Surveillance System (ISS), the Intelligent Transport System (ITS), robotics, and automotive safety. However, the human body′s position, angle, and dress in a video scene are complicated and changeable, which have a great influence on the detection accuracy. Through the analysis on the pros and cons of Census Transform Histogram (CENTRIST), a novel feature was presented for human detection-Ternary CENTRIST (T-CENTRIST). The T-CENTRIST feature took the relationship between each pixel and its neighborhood pixels into account. Meanwhile, the relevancy among these neighborhood pixels was also considered. Therefore, the proposed feature description method can reflect the silhouette of pedestrian more adequately and accurately than that of CENTRIST. Secondly, a fast pedestrian detection framework based on T-CENTRIST in infrared image was proposed, which introduced the idea of extended blocks and the integral image. Finally, experimental results verified the effectiveness of the proposed pedestrian detection method.