基于迁移学习技术的隧道围岩岩性识别方法

为实现隧道围岩岩性的自动识别与分类,提出了基于迁移学习技术的围岩岩性识别方法。首先,通过采用Inception-ResNet-V2(IRV2)卷积神经网络模型在Image-Net数据集上进行预训练,并利用模型迁移学习技术对岩石图片数据集(包含花岗岩、石灰岩、玄武岩和页岩)进行再训练,获取隧道围岩岩性识别模型;然后,对IRV2进行模型测试,并与ResNet-50、Inception-V3和VGG16三种模型的识别性能进行对比;最后,进行子图像法与整体图像法的识别效果对比试验。实验结果表明:(1)IRV2的各项分类性能指标均表现为最优,且均可达到90%以上,表明该模型可以实现围岩岩性的有效识别与精确分类;(2)对于具有更加突出的纹理、结构和构造等外部特征的岩石图片,模型的识别性能更好;(3)子图像法相比于整体图像法可有效提高模型的识别性能。

Accelerating SAGE algorithm in PET image reconstruction by rescaled block-iterative method

A new method to accelerate the convergent rate of the space-alternatinggeneralized expectation-maximization (SAGE) algorithm is proposed. The new rescaled block-iterativeSAGE (RBI-SAGE) algorithm combines the RBI algorithm with the SAGE algorithm for PET imagereconstruction. In the new approach, the projection data is partitioned into disjoint blocks; eachiteration step involves only one of these blocks. SAGE updates the parameters sequentially in eachblock. In experiments, the RBI-SAGE algorithm and classical SAGE algorithm are compared in theapplication on positron emission tomography (PET) image reconstruction. Simulation results show thatRBI-SAGE has better performance than SAGE in both convergence and image quality.

Experimental Study on the Flow Around Two Tandem Cylinders with Unequal Diameters

In this paper, flow around two circular cylinders in tandem arrangement with unequal diameters has been investigated using the particle image velocimetry technique(PIV) in a water channel. The upstream to downstream diameter ratio was kept constant at d/D = 2/3, the centre-to-centre distance was varied from 1.2D to 5D and the Reynolds number was varied from 1200 to 4800. The flow characteristics were analyzed through ensemble-averaged patterns of velocity, vorticity, normalized Reynolds stress contours and streamlines. Based on ensemble-averaged and instantaneous flow fields, different flow patterns, including single-wakeshedding at small spacing ratio, bi-stable flow behavior(alternating behavior of reattachment and vortex shedding) at intermediate spacing ratio and co-shedding pattern at large spacing ratio were observed. The effects of Reynolds number and the centre-to-centre spacing ratio on flow patterns and turbulent characteristics were also investigated. It was found that the diameter ratio appears to have a certain effect on the flow patterns at intermediate spacing ratios, where the reattachment of shear layer depends on the lateral width of the wake flow in the lee of the upstream cylinder. Extensive discussion on the distributions of Reynolds stress and turbulent kinetic energy was presented.

Flow Field Around Rushton Turbine in Stirred Tank by Particle Image Velocimetry Measurement

In this paper, particle image velocimetry (PIV) was used to measure the mean and root meansquare(RMS) velocity in the stirred tank with six-flat blade Rushton turbine and with no baffles. Two typesof motion patterns were studied. One was that the impeller runs at constant speed, the other was that the impellerruns at time-dependent speed and in a periodic way. The emphasis of the paper was on the comparison of meanand RMS velocity vector maps and profiles between these two types of motion patterns, and especial attention waspaid to the comparison of the mean velocity, time-averaged RMS velocity, phase averaged RMS velocity betweenthe constant 3 RPS (revolution per second) and time-dependent operation. The Reynolds number was between 763and 1527. The study explained the mechanism that time-dependent RPS is more efficient for mixing than that ofconstant RPS.

Improving Image Copy-Move Forgery Detection with Particle Swarm Optimization Techniques

Copy-Move Forgery(CMF) is one of the simple and effective operations to create forged digital images.Recently,techniques based on Scale Invariant Features Transform(SIFT) are widely used to detect CMF.Various approaches under the SIFT-based framework are the most acceptable ways to CMF detection due to their robust performance.However,for some CMF images,these approaches cannot produce satisfactory detection results.For instance,the number of the matched keypoints may be too less to prove an image to be a CMF image or to generate an accurate result.Sometimes these approaches may even produce error results.According to our observations,one of the reasons is that detection results produced by the SIFT-based framework depend highly on parameters whose values are often determined with experiences.These values are only applicable to a few images,which limits their application.To solve the problem,a novel approach named as CMF Detection with Particle Swarm Optimization(CMFDPSO) is proposed in this paper.CMFD-PSO integrates the Particle Swarm Optimization(PSO) algorithm into the SIFT-based framework.It utilizes the PSO algorithm to generate customized parameter values for images,which are used for CMF detection under the SIFT-based framework.Experimental results show that CMFD-PSO has good performance.

New learning subspace method for image feature extraction

A new method of Windows Minimum/Maximum Module Learning Subspace Algorithm(WMMLSA) for image feature extraction is presented.The WMMLSM is insensitive to the order of the training samples and can regulate effectively the radical vectors of an image feature subspace through selecting the study samples for subspace iterative learning algorithm,so it can improve the robustness and generalization capacity of a pattern subspace and enhance the recognition rate of a classifier.At the same time,a pattern subspace is built by the PCA method.The classifier based on WMMLSM is successfully applied to recognize the pressed characters on the gray-scale images.The results indicate that the correct recognition rate on WMMLSM is higher than that on Average Learning Subspace Method,and that the training speed and the classification speed are both improved.The new method is more applicable and efficient.

Determination of pile failure mechanism under pullout test in loose sand

Estimating the deformation of soil around the pile contributes to reliable design of structures under pullout force. This work presents the results of a series of small-scale physical modelling tests designed to investigate the uplift resistance of piles with diameter of 5 cm and slenderness ratios of 1, 2, 3 and 4 in loose sand. Close photogrammetric technique and particle image velocimetry(PIV) were employed to observe the failure patterns due to uplift force on piles. The results show that the shear zones curve slightly outward near the ground surface. After peak resistance, the shear strain concentrates into a pair of narrow shear bands,then a flow around mechanism is formed accompanied by a reduction in the uplift resistance. The results from the laboratory tests were verified by analytical method proposed by Chattopadhyay and PLAXIS 2D and 3D finite element method software. It is found that the depth and width of the failure surface increase with the increment of the slenderness ratio. A good agreement is observed among the measured bearing capacity and obtained failure surface of the models and the results of numerical modelling. Finally, the maximum deformation of loose and dense sand respectively with densities of 25% and 75% were compared in the stage of fully removing pile. The results shows that the deformation of the soil is related to its density, therefore it depends on its dilatancy.

Droplet Contact Angle Measurement on Microstructures of Octocoral Sclerites-ESEM Image Analysis

This work explores the methodology for micron-scale water droplet contact angle derivation for the warty surface of octocoral sclerites. The calcite-made sclerites of the Red Sea octocoral Dendronephthya hemprichi have been chosen as a model for this study. Water droplet condensation on the sclerites has been in-situ investigated using Quanta 200 FEG （field emission gun） ESEM （environmental scanning electron microscope） under wet environmental conditions. Two different analysis methods of droplet top and side views have been applied to determine the contact angle based on the secondary electron images. The ESEM image analysis for the sclerites indicates that their surface is hydrophilic. The microscopic contact angle is measured to be 45.3°±6.3°. The macroscopic contact angle has been calculated by using the Wenzel model for the surface texturing of the sclerites.

Effects of solution treatment on mechanical properties and microstructures of Al-Li-Cu-Mg-Ag alloy

Mechanical properties and microstructures of Al-Li-Cu-Mg-Ag alloy after solution treatments were investigated by means of optical microscopy （OM）, tensile test, hardness measurement and electrical conductivity test, differential scanning calorimetric （DSC）, energy dispersive X-ray （EDX）, scanning electron microscopy （SEM） and transition electron microscopy （TEM）, respectively The results show that both tensile strength and hardness increase first and then decrease with temperature at constant holding time of 30 min with maximum strength and hardness appearing at 520 ℃. Tensile strength, hardness and elongation of samples treated at 520 ℃ for 30 min are 566 MPa （σb）, 512 MPa （σ0.2）, HB 148 and 8.23% （δ）, respectively. There are certain amount of fine T1 （AI2CuLi） phase dispersing among AI substrates according to TEM images. This may result in mixed fracture morphology with trans-granular and inter-granular delamination cracks observed in SEM images.

Research on Image Segmentation Algorithm based on Fuzzy C-mean Clustering

This paper presents a fuzzy C- means clustering image segmentation algorithm based on particle swarm optimization, the method utilizes the strong search ability of particle swarm clustering search center. Because the search clustering center has small amount of calculation according to density, so it can greatly improve the calculation speed of fuzzy C- means algorithm. The experimental results show that, this method can make the fuzzy clustering to obviously improve the speed, so it can achieve fast image segmentation.

Automatic recognition and quantitative analysis of Ω phases in Al-Cu-Mg-Ag alloy

The main methods of the second phase quantitative analysis in current material science researches are manual recognition and extracting by using software such as Image Tool and Nano Measurer. The weaknesses such as high labor intensity and low accuracy statistic results exist in these methods. In order to overcome the shortcomings of the current methods, the Ω phase in A1-Cu-Mg-Ag alloy is taken as the research object and an algorithm based on the digital image processing and pattern recognition is proposed and implemented to do the A1 alloy TEM （transmission electron microscope） digital images process and recognize and extract the information of the second phase in the result image automatically. The top-hat transformation of the mathematical morphology, as well as several imaging processing technologies has been used in the proposed algorithm. Thereinto, top-hat transformation is used for elimination of asymmetric illumination and doing Multi-layer filtering to segment Ω phase in the TEM image. The testing results are satisfied, which indicate that the Ω phase with unclear boundary or small size can be recognized by using this method. The omission of these two kinds of Ω phase can be avoided or significantly reduced. More Ω phases would be recognized （growing rate minimum to 2% and maximum to 400% in samples）, accuracy of recognition and statistics results would be greatly improved by using this method. And the manual error can be eliminated. The procedure recognizing and making quantitative analysis of information in this method is automatically completed by the software. It can process one image, including recognition and quantitative analysis in 30 min, but the manual method such as using Image Tool or Nano Measurer need 2 h or more. The labor intensity is effectively reduced and the working efficiency is greatly improved.

Embedding ensemble tracking in a stochastic framework for robust object tracking

We propose an algorithm of embedding ensemble tracking in a stochastic framework to achieve robust tracking performance under partial occlusion,illumination changes,and abrupt motion.It operates on likelihood images generated by the ensemble method,and combines mean shift and particle filtering in a principled way,where a better proposal distribution is de-signed by first propagating particles via a motion model,and then running mean shift to move towards their local peaks in the likelihood image.An observation model in the particle filter incorporates global and local information within a region,and an adaptive motion model is adopted to depict the evolution of the object state.The algorithm needs fewer particles to manage the tracking task compared with the general particle filter,and recaptures the object quickly after occlusion occurs.Experiments on two image sequences demonstrate the effectiveness and robustness of the proposed algorithm.

Fourier mode decomposition of PIV data.

The modal decomposition technique is one of the most effective methods for studying the flow dynamics in a complex flow. By rejuvenating the discrete Fourier transform(DFT), this paper proposes a Fourier mode decomposition(FMD) method for the time series of particle image velocimetry(PIV) data from the fluid field. An experimental case concerning the control of the flow around a circular cylinder by a synthetic jet positioned at the rear stagnation point is used to demonstrate the use of the FMD method. In the three different regimes where the natural shedding frequency and actuation frequency dominate respectively or simultaneously, it is found that the FMD method is capable of extracting the dynamic mode along with its amplitude and phase according to the selected characteristic frequency based on the global power spectrum. For the quasiperiodic flow phenomena presented in this particular case, the FMD method can reconstruct the original flow field using the zero-th mode and the selected mode corresponding to the characteristic frequency. Similarities and differences between the FMD method and the dynamical mode decomposition(DMD) and proper orthogonal decomposition(POD) methods are also discussed.

A hyperspectral image endmember extraction algorithm based on generalized morphology

Generalized morphological operator can generate less statistical bias in the output than classical morphological operator. Comprehensive utilization of spectral and spatial information of pixels, an endmember extraction algorithm based on generalized morphology is proposed. For the limitations of morphological operator in the pixel arrangement rule and replacement criteria, the reference pixel is introduced. In order to avoid the cross substitution phenomenon at the boundary of different object categories in the image, an endmember is extracted by calculating the generalized opening-closing(GOC) operator which uses the modified energy function as a distance measure. The algorithm is verified by using simulated data and real data. Experimental results show that the proposed algorithm can extract endmember automatically without prior knowledge and achieve relatively high extraction accuracy.

Preparation of nano-porous AlN micro-rods

In this paper, nano-porous A1N micro-rods were prepared and characterized by X-ray diffraction （XRD）, scanning electron microscopy （SEM）, transmission electron microscopy （TEM）, Brunauer-Emett-Teller （BET） method of nitrogen adsorption at low temperature, UV-via absorption spectrometry and Raman spectrometry. The SEM images and TEM images showed that the A1N samples were micro-size rods with pores in their surfaces evenly. The length of the rod was in the range of a few mi- crons to tens of micron, the diameter of the rods was about one micron, the diameter of the pore was about hundreds of nano- meter and the wall thickness of the pore about tens of nanometer. The BET surface area of the sample was 41.424 ma/g. The optical spectra showed that the optical properties of the A1N samples almost agreed with that of A1N bulk or film.