Abstract:Aiming at the demands of time-sensitive building health monitoring to promote the automation level of surface disease visual inspection， scene information guided UAV inspection mission planning was proposed. Based on the scene’s prior information， two observation modes， parallel observation and envelope observation， were designed for the structural characteristics of the building， which realized the full coverage obstacle avoidance inspection of the individual building in the narrow space as well as the observation of whole building with millimeter resolution. Meanwhile， a series of effective quantitative indexes for the overall evaluation of the inspection quality were put forward. The facade of the building was divided into 3 720 subregions. The surface cracks were identified and classified by a deep residual network. The result shows that 13 wrong subregions and 14 missing subregions reflect the high accuracy of crack identification. The crack skeletons are mapped to the reconstructed 3D model， which provides data support for the integrated expression of crack morphology and building information. This study combines high-precision 3D reconstruction with surface disease recognition， providing a practical observation and analysis method for integrated building health monitoring.

Abstract:Compact wearable laser scanning system has the advantages of low cost， small volume and easy operation， which provides a feasible strategy for indoor and outdoor seamless 3D geographic information acquisition. It is an effective supplement to the traditional mobile measurement system. Compared with the existing handheld and backpack laser scanning system， the helmet-based platform keeps the same direction with the user's sight， with the characteristics of “What you see is what you get”， liberating user's hands. Using the self-designed helmet-based laser scanning system， named "WHU-Helmet"， this paper proposes a multi-scale NDT based LiDAR-IMU SLAM. Taking two typical indoor and outdoor environments as the study area （office building and subway station）， the 3D point clouds obtained by "WHU-Helmet" are compared with the 3D point clouds obtained by terrestrial laser scanning. The average error is less than 0.44 m and the root mean square error is less than 0.23 m， which demonstrated that the helmet-based laser scanning system has great potential in indoor and outdoor seamless 3D mapping.

Abstract:Water depth is an important topographical parameter that characterizes shallow ocean and coastal environment. Photon-counting light detection and ranging （LiDAR） can penetrate a certain depth of water and provide reliable date support for water depth information extraction. Taking the islands and reefs in the South China Sea as an example， this paper uses the only in-orbit spaceborne photon-counting LiDAR - Ice， Cloud and land Elevation Satellite-2/Advanced Topographic Laser Altimeter System （ICESat-2/ATLAS） to carry out research on depth extraction and accuracy evaluation in shallow water areas of islands and reefs. Firstly， confidence parameter is adopted to remove the coarse noise photons， and the photons in the water surface and bottom are separated according to their density. Secondly， interval estimate and the modified ordering points to identify the clustering structure （OPTICS） are utilized to filter out noise photons in the water surface and bottom respectively， and the modified OPTICS is changed twice by the filter parameters. Then the water surface elevation is obtained by using the random sample consensus （RANSAC） algorithm. Thirdly， the bathymetric information is achieved by the refraction and tide correction. Finally， the airborne bathymetric LiDAR data of South China Sea is used to validate and evaluate the bathymetric accuracy. Compared with the depth results extracted from the high confidence photons and adaptive variable ellipse filtering bathymetric method （AVEBM）， the proposed noise removal algorithm has a higher F value， which is increased by 5.87% and 3.38% respectively. The experimental results indicated that the R² of bathymetric results obtained by ATLAS and airborne LiDAR is 0.91 and root mean square error （RMSE） is 0.53m.

Abstract:This paper presents a tree diameter at breast height （DBH）estimation algorithm using terrestrial laser scanning （TLS） based on the randomized random sample consensus （RANSAC） model. First， the forest cloud data were filtered by the cloth simulation filtering （CSF） algorithm. Digital elevation model （DEM） and the point cloud at the DBH of the tree were extracted. Then， Euclidean distance clustering was performed. Finally， the tree model was fitted on the basis of the randomized RANSAC model. The algorithm was verified on laser point cloud data from two sample plots in Qingpu District， Shanghai. The average biases from the actual diameter of trees were 0.79 cm and 0.52 cm， respectively. Experiment results show that the algorithm is better than those based on Hough transform or the least-squares in terms of the accuracy and running time.

Abstract:Reconstructing historical land cover dynamics by time series remote sensing images is the basis of geographic scene sensing and modeling. However， classification of historical archived remote sensing images is quite difficult due to the limited training samples and low interpretation capability for multi-temporal images. A spatio-temporal training sample transfer method from existing land cover products and corresponding remote sensing images was proposed based on geometric and attribute constrains. The transferred training samples were then embedded into a multi-temporal land cover classification framework to solve the problem of high sample labeling cost in multi-temporal image classification. The multi-temporal classification results in the Taihu Basin show that the proposed unsupervised sample transfer method can make full use of the prior geometric constraints and spectral properties of the products and images， and the classification accuracy in large scale area is over 89%. The results demonstrate that the proposed method is effective for land cover updating and geographic environment evolution modelling.

Abstract:Based on China land cover dataset (CLCD) and contour extraction algorithm，this paper systematically studies the form and area expansion of rural settlements architecture in China from 1990 to 2020. The results show that in the process of rapid urbanization， the rural population in China is decreasing while the rural settlements are expanding continuously. In 2020， the area of rural settlements in China was about twice that of 1990. Small rural settlements have a faster expansion rate and gradually slow down after a specific threshold. The expansion rate of rural settlements has obvious spatial difference， and the expansion rate in the western China is faster than that in the eastern China. The continuous expansion of rural settlements reflects the encouraging progress in China’s balanced development between urban and rural areas. The “double growth” of urban and rural construction land may have a negative impact on agricultural and ecological land， which should be paid great attention to.

Abstract:A topology optimization method of multi-material structure based on deep convolution neural network （CNN） is proposed， which can predict the optimized structure of multi-material in a very short time without any iteration. The popular U-Net network structure is adopted to improve the edge extraction ability of neural network. To train the network， the ordered multi-material SIMP （isotropic real material penalty density method） interpolation method （Ordered SIMP） is used to generate multi-material optimal structure data sets under random loading conditions， mass fraction and cost fraction. The efficiency and accuracy of the proposed method are compared with traditional algorithms， and the performance of the proposed method is evaluated. The results show that the proposed method can significantly reduce the computational cost with little sacrifice on the performance of the design scheme. The proposed method has great potential and broad application prospects for topology optimization in the practice of multi-material structure design in the future.

Abstract:Wind field of downburst was simulated in the atmospheric boundary layer wind tunnel by a self-made generator， and synchronous pressure measurements on the internal and external pressures of a typical high-rise building with an opening were conducted. The characteristics of internal pressure and net pressure as well as the internal pressure gain function were investigated in the terrain catergories downburst and B. Furthermore， orifice parameters were identified， and the theoretical calculation results of peak internal pressure were compared with the corresponding experimental results. Results show that both the mean and peak internal pressure coefficients decrease first and then increase with the increase of wind direction angle. Peak net pressures on both side-walls are the largest， which will easily lead to damage of side-walls. However， net pressure on the front wall with opening is the smallest due to the positive correlation between internal and external pressures. Besides， the amplitudes of internal pressure gain function in the wind direction of 50° are much larger than those in other wind directions， and the internal pressure energy at the Helmholtz resonance frequency is also significantly larger than that at other frequencies. Except for the wind direction angles of 70°~100°， the theoretical values of peak internal pressure coefficient are in good agreement with the experimental values.

Abstract:In order to study the damage mechanism and flexural performance of lithium slag concrete beams under coupling between loading and thermal-cold cycling，flexural loading tests were made on 4 groups of lithium slag concrete beams， which underwent the combined effect of a series of thermal-cold cycling （0， 100， 300） and static load levels （0， 0.2， 0.35）. In addition，a constitutive model of lithium slag concrete was established on the basis of the axial compression test. Through the force analysis of the beam section， the calculation formula of the flexural bearing capacity of the lithium slag concrete beam was proposed. A comparative study was made of the calculation and experiment results， and the theoretical calculation results prove to be less than those of the experiments. From the prospect of the safety-reservation，the study provides a reference for the calculation of the flexural bearing capacity of lithium slag concrete rectangular cross-section beams in the areas of large temperature changes.

Abstract:PFC2D software was used to simulate sand samples prepared by moist tamping method and deposition method. Undrained cyclic loading tests were performed on samples to analyze the microstructure changes during the dynamic loading process. The test results indicated that the long axis direction of sand particles tended to distribute horizontally on the compression side but vertically on the extension side when cyclic loading was applied. Although the distribution of the normal directions of particle contacts was contrary to the long axis direction， both anisotropy coefficients were increased gradually with the increasing of cyclic number. Besides， the coordination number of the samples decreased and finally fluctuated within the range of 0-3 after liquefaction. The samples obtained from moist tamping method exhibited less anisotropy than the deposition samples. Thus， the former one required more loading cycles to reach liquefaction， which implies the layered samples owned higher resistance to liquefaction. The differences in liquefaction resistance between two samples were more obvious for dense sand than medium-dense sand.

Abstract:In order to study the response process of storm surge on the beach， video monitoring and computer vision were used to obtain color images of the beach through several cameras set up on the shore high place. The improved Canny algorithm combined with dual threshold processing was used to automatically extract the shoreline. Used camera calibration to obtain the geographic coordinates of the shoreline. The multi-temporal shorelines were interpolated to obtain the shoreline coordinates and beach topography. Through continuous monitoring and analysis， it was found that Pingshuiqiao beach in Qinhuangdao responded sharply to the storm surge process in a short time. Sediment was transported westward and toward the sea， the beach berm suffered obviously erosion and sand was coarsened. This caused the intertidal siltation， the beach slope decrease， and the coastline advancing 4.34 m toward the sea. In the recovery period after the storm surge， the high tide zone of Pingshuiqiao beach headland shield section was slightly eroded and silted in the middle and low tide zone， and the sediment was transported offshore. The evolution of the straight section is mainly affected by the small and medium recovery waves， which show the trend of upper siltation and lower erosion， and the sediment was transported from the middle and low tide zone to the high tide zone.

Abstract:Based on the engineering background that the vehicle exhaust emission forms oil film on the surface of tunnel asphalt pavement， which leads to the rapid attenuation of pavement skid-resistance performance， the attenuation law of asphalt mixture skid-resistance performance under the combined action of humidity and tail gas and the formation mechanism of tail gas oil film are explored. Using constant temperature and humidity curing box and self-made exhaust gas treatment system， the combined effect of different humidity and tail gas on three kinds of asphalt mixture rut plate specimens was simulated， and the British pendulum number （BPN） was used to characterize the skid-resistance performance of the specimens. In addition， molecular dynamics simulation was carried out on the water （anhydrous） tail gas - asphalt interface to explore the micro mechanism of tail gas oil film formation. The results show that the British pendulum number of three asphalt mixtures shows a linear downward trend with the increase of tail gas treatment time under relatively dry and completely moist conditions. After tail gas treatment for 1h under completely moist conditions， the attenuation range of skid-resistance performance of OGFC mixture is nearly 30% lower than that of the other two mixtures. OGFC mixture has better skid-resistance performance， and has good skid-resistance durability under the combined action of humidity and tail gas. The skid-resistance performance of asphalt mixture caused by tail gas oil film is difficult to dissipate naturally and has cumulative effect. Tail gas oil film is the main reason for the rapid attenuation of skid-resistance performance of asphalt mixtures. The tail gas oil film is mainly formed in the process of cooling after the tail gas is discharged. Water molecules mainly exist in the range of 0~5 ? near the tail gas molecules， filling the gap between tail gas molecules. Its strong polarity can attract tail gas molecules and contribute to the adsorption of tail gas on the surface of asphalt molecules.

Abstract:The research progress of health management technology for vehicle IGBT module was reviewed from four aspects： failure mechanism， condition monitoring， fault diagnosis， damage assessment and life prediction. The key problems and technical advantages and disadvantages of the current research were pointed out. Moreover， the future development trend of health management technology for vehicle IGBT module was prospected.

Abstract:Reasonable scaled model of high-speed train is very important for aerodynamic evaluation and research in wind tunnel. Based on the computational fluid dynamics method， the selection method of the shortest tail model and the reasonable scale model of the head car in wind tunnel test was studied. The results reveals that： the drag coefficient of the head car increases with the shortening of the tail model， and the negative pressure of the tail moves upstream， which reduces the pressure at the inter-coach space. The deviation between the drag of the head car +0.4L tail model and that of the three cars model is 2.8%， which can be used as the shortest model of the head car. The selection of the scale model of the head car is affected by the wind tunnel boundary effect， Reynolds number effect and ground effect. The Reynolds number effect reduces the frictional drag， while the ground effect increase the pressure drag. The 1：8， 1：4 scale model and full-scale model's head car pressure drag deviation are 12.7% and 7.2%， which account for 63.9%， 67.2% and 72.5% of the head car's total drag respectively. The result shows that， it’s more reasonable to choose the 1：4 scale model as high-speed train experiment model in open-section wind tunnel.

Abstract:The directed relation graph with error propagation （DRGEP）， full species sensitivity analysis （FSSA） and directed relation graph with path flux analysis （DRGPFA） were used to reduce the mechanism of methanol. Based on the reduced mechanism， the key primitive reactions were found by sensitivity analysis， then their pre-exponential factors were adjusted by coarse perturbation and refinement perturbation analysis which were on this basis of pre-factor perturbation method， to obtain the optimal optimization mechanism and verify it. The results show that the FSSA reduced mechanism has 16 species and 65 reactions， which can accurately predict ignition delay time and laminar flame velocity. The prediction accuracy of the V6-3-4 optimized mechanism is better than that of the FSSA reduced mechanism， and the predicted value is much closer to the experimental value than the detailed mechanism.

Abstract:Bohai Bay Basin is an important oil-bearing basin in eastern China. The discovery of Bozhong 19-6 gas field， the largest integrated condensate gas field in eastern China in 2018， has opened up a new field of deep gas exploration in Bohai bay Basin， showing great potential of gas exploration in oil-type basin. The Bozhong 19-6 structure belt has a unique distribution characteristic of oil accumulation in shallow part and gas accumulation in deep part. The hydrocarbon a ccumulation period and adjustment process in buried hill reservoir was investigated using a suite of hydrocarbon generation and expulsion history， tectonic thermal evolution history， magmatic activity， characteristics of fluid inclusions， microfluorescence determination and temperature measurement technology by taking Bozhong 19-6 structure belt as an example. The results show that ： The inclusions in the condensate field in the Bozhong 19-6 structural belt have similar distribution ranges of salinity and homogenization temperature at different depths， and the main peak temperature of homogenization temperature at different depths is higher than that of the stratum temperature， which indicates that the oil and gas charging belongs to the event accumulation， which has the characteristics of layer penetration， instantaneous and high temperature.The magmatic activity developed in the Bozhong 19-6 tectonic belt. The magmatic intrusion accelerated the maturation of organic matter and promoted the hydrocarbon generation and expulsion of source rocks， and the source rocks formed a lot of condensate oil and gas in a short time.Bozhong 19-6 buried structure belt， the early formation of the ancient reservoir is affected by the new tectonic movement， adjusted upward for Bozhong 19-4 oilfield， magmatic upwelling activity， with the mantle-derived CO₂ supercritical state of liquid， speeding up the process of oil and gas migration， condensate oil and gas rapidly filling into Bozhong 19-6 buried hill reservoir， forming Bozhong 19-6 gas field.

Abstract:Based on the idea of ??TOPSIS （Technique for Order Preference by Similarity to Ideal Solution）， this research builds a multi-objective programming model to solve the optimal combination weights. It conducts an empirical analysis on the output of 198 patents on 18 R&D and transformation functional platforms in Shanghai， and finds that， the combination empowering model constructed in this study has a high identification capability of patent transformation risk （Z=5.546）， which can determine the actual R&D efficiency of the platform more scientifically. Research innovations are manifested in two aspects： The multi-objective nonlinear programming model is constructed by positive and negative ideal points to carry out combination weighting， which avoids the disadvantages of the existing research， such as a large number of overlapping evaluation scores of patent failure conversion and successful transformation， and low distinguishing ability of the two types of transformation； Through the test of whether the efficiency score of failed transformation is significantly less than the median of the efficiency score of successful conversion， the rationality of efficiency evaluation model is verified， and the defect that existing research neglects to verify the rationality of platform patent conversion efficiency measurement model is avoided.