基于改进的ResNet34网络模型的苹果叶病害识别

针对目前自然环境下苹果叶病害图像识别准确率不高和模型参数量多的问题,在深度残差网络ResNet34的基础上,提出了一种改进模型。首先对原始ResNet34第一层卷积层进行卷积核的替换,将其中的大卷积核改为小卷积核;进而对残差结构进行优化,将残差结构卷积层中的两个大小为3×3卷积核替换成大小为1×1、3×3和1×1串联的卷积核;改变残差网络的block数,由原始的[3,4,6,3]改为[1,2,12,1];对于虚线残差结构,通过添加一个平均池化层来进行下采样操作,取消了1×1卷积层的下采样功能。实验结果表明,改进的ResNet34模型在苹果叶病害数据集分类任务中获得了96.7%的识别准确率,平均F1分数达到了96.7%。与原始ResNet34模型相比,改进的ResNet34模型在识别能力得到小幅度提升的同时,模型参数量降低了43.3%。

Effect of gas saturation on P-wave velocity in tight sandstone

By measuring the variation of the P-and S-wave velocities of tight sandstone samples under water saturation,it was confirmed that with the decrease in water saturation,the P-wave velocity first decreased and then increased.The variation in velocity was influenced by the sandstone’s porosity.The commonly used Gassmann equation based on fluid substitution theory was studied.Comparing the calculated results with the measured data,it was found that the Gassmann equation agreed well with the measured data at high water saturation,but it could not explain the bending phenomenon of P-wave velocity at low saturation.This indicated that these equations could not accurately describe the relationship between fluid content and rock acoustic velocity.The reasons for this phenomenon were discussed through Taylor’s expansion.The coefficients of the fitting formula were calculated and verified by fitting the measured acoustic velocity changes of the cores.The relationship between P-wave velocity and saturation was discussed,which provides experimental support for calculating saturation using seismic and acoustic logging data.

The gas discharge visualization(GDV)order parameter model based on the principle of mastering both permanence and change

Objective To investigate the human body’s complex system,and classify and characterize the human body’s health states with“a comprehensive integrated method from qualitative to quantitative”.Methods This paper introduces the concept of“order parameters”and proposes a method for establishing an order parameter model of gas discharge visualization(GDV)based on the principle of“mastering both permanence and change(MBPC)”.The method involved the fol-lowing three steps.First,average luminous intensity(I)and average area(S)of the GDV im-ages were calculated to construct the phase space,and the score of the health questionnaire was calculated as the health deviation index(H).Second,the k-means++clustering method was employed to identify subclasses with the same health characteristics based on the data samples,and to statistically determine the symptom-specific frequencies of the subclasses.Third,the distance(d)between each sample and the“ideal health state”,which determined in the phase space of each subclass,was calculated as an order parameter describing the health imbalance,and a linear mapping was established between the d and the H.Further,the health implications of GDV signals were explored by analyzing subclass symptom profiles.We also compare the mean square error(MSE)with classification methods based on age,gen-der,and body mass index(BMI)indices to verify that the phase space possesses the ability to portray the health status of the human body.Results This study preliminarily tested the reliability of the order parameter model on data samples provided by 20 participants.Based on the discovered linear law,the current model can use d calculated by measuring the GDV signal to predict H(R^(2)>0.77).Combined with the symptom profiles of the subclasses,we explain the classification basis of the phase space based on the pattern identification.Compared with common classification methods based on age,gender,BMI,etc.,the MSE of phase space-based classification was reduced by an order of magnitude.Conclusion In this study,the GDV order parameter model based on MBPC can identify sub-classes and characterize individual health levels,and explore the TCM health meanings of the GDV signals by using subjective-objective methods,which holds significance for establishing mathematical models from TCM diagnosis principles to interpret human body signals.

基于多尺度时序特征的信号调制样式识别算法

在认知无线电应用中,当前基于深度学习的信号调制样式识别算法存在运算效率低、复杂度较高的问题,为此本文提出了一种基于多尺度时序特征的信号调制识别算法。该算法首先利用多层卷积层提取不同尺度的时序数据,将数据进行特征融合后使用长短期记忆网络提取时序特征,最后由输出层输出识别结果,并通过网络结构的设计和优化,降低了算法复杂。实验结果表明,在包含11种调制信号的原始I/Q信号测试集上,在信噪比为4 dB及以上时,该算法识别准确率达到90%以上。与同等识别准确率的算法相比,该算法的复杂度更低,在嵌入式设备Jetson Nano和树莓派4B上的推理时间更短,具有更好的工程应用价值。

Estimation of the van Genuchten Soil Water Retention Properties from Soil Textural Data

The van Genuchten (vG) function is often used to describe the soil water retention curve (SWRC) of unsaturated soils and fractured rock. The objective of this study was to develop a method to determine the vG model parameter m from the fractal dimension. We compared two approaches previously proposed by van Genuchten and Lenhard et al. for estimating m from the pore size distribution index of the Brooks and Corey (BC) model. In both approaches we used a relationship between the pore size distribution index of the BC model and the fractal dimension of the SWRC. A dataset containing 75 samples from the UNSODA unsaturated soil hydraulic database was used to evaluate the two approaches. The statistical parameters showed that the approach by Lenhard et al. provided better estimates of the parameter m. Another dataset containing 72 samples from the literature was used to validate Lenhard's approach in which the SWRC fractal dimension was estimated from the clay content. The estimated SWRC of the second dataset was compared with those obtained with the Rosetta model using sand, silt, and clay contents. Root mean square error values of the proposed fractal approach and Rosetta were 0.081 and 0.136, respectively, indicating that the proposed fractal approach performed better than the Rosetta model.

Linear Support Vector Machine Based Inter-harmonic Modeling and Parameter Estimation

There are often system. The high measure many inter-harmonics in power t accuracy of inter-harmonics order, amplitude and initial phase is needed. A new approach is presented for inter-harmonic modeling and parameter estimation based on linear support vector machine （SVM）. Firstly, parameter estimation of linear model is realized based on standard linear SVM. Then, interharmonic model is transformed to a linear model according to trigonometric functions. The approach obtains order of inter-harmonic model with windowed Blackman-Tukey （BT） spectrum analysis, and gets number and frequency of harmonics. Finally, the linear SVM is applied to estimate the inter-harmonic parameters, amplitude and initial phase. The simulation results show that the proposed approach has high precision and good antinoise. The accuracy of three parameters are all higher than 98%.

Early warning model for slope debris flow initiation

Early warning model of debris flow is important for providing local residents with reliable and accurate warning information to escape from debris flow hazards. This research studied the debris flow initiation in the Yindongzi gully in Dujiangyan City, Sichuan province, China with scaled-down model experiments. We set rainfall intensity and slope angle as dominating parameters and carried out 20 scaled-down model tests under artificial rainfall conditions. The experiments set four slope angles(32°, 34°, 37°, 42°) and five rainfall intensities(60 mm/h, 90 mm/h, 120 mm/h, 150 mm/h, and 180 mm/h) treatments. The characteristic variables in the experiments, such as, rainfall duration, pore water pressure, moisture content, surface inclination, and volume were monitored. The experimental results revealed the failure mode of loose slope material and the process of slope debris flow initiation, as well as the relationship between the surface deformation and the physical parameters of experimental model. A traditional rainfall intensity-duration early warning model(I-D model) was firstly established by using a mathematical regression analysis, and it was then improved into ISD model and ISM model(Here, I is rainfall Intensity, S is Slope angle, D is rainfall Duration, and M is Moisture content). The warning model can provide reliable early warning of slope debris flow initiation.

Retrieval algorithm of sea surface wind vectors for WindSat based on a simple forward model

WindSat/Coriolis is the first satellite-borne polarimetric microwave radiometer, which aims to improve the potential of polarimetric microwave radiometry for measuring sea surface wind vectors from space. In this paper, a wind vector retrieval algorithm based on a novel and simple forward model was developed for WindSat. The retrieval algorithm of sea surface wind speed was developed using multiple linear regression based on the simulation dataset of the novel forward model. Sea surface wind directions that minimize the difference between simulated and measured values of the third and fourth Stokes parameters were found using maximum likelihood estimation, by which a group of ambiguous wind directions was obtained. A median filter was then used to remove ambiguity of wind direction. Evaluated with sea surface wind speed and direction data from the U.S. National Data Buoy Center （NDBC）, root mean square errors are 1.2 rn/s and 30~ for retrieved wind speed and wind direction, respectively. The evaluation results suggest that the simple forward model and the retrieval algorithm are practicable for near-real time applications, without reducing accuracy.

Correlates of subjectively and objectively measured physical activity in young adolescents

Background： Many studies examining individual-level correlates in youth utilize self-report rather than objective measures of physical activity （PA）. This utilization of self-report may result in associations that are not present when examining objectively measured PA. The present study investigates the relationship between hypothesized correlates of PA with objectively and subjectively measured PA. Methods： Participating children （n 232, 101 males, mean age - 12.3 years） provided a minimum of four monitored days of PA （via accelerometer） and completed a survey assessing moderate-to-vigorous physical activity （MVPA）, sport competence, appearance, enjoyment, and self-efficacy. Height and weight were measured and body mass index （BMI） was calculated. Results： Hierarchical regression models controlling for sex, race, and BMI Z-score showed that only sex and BMI Z-score were significant correlates of objective MVPA while only sex was a significant correlate of objective total PA. However, in a separate model examining the relationship with subjective MVPA, enjoyment of PA and self-efficacy for PA were the only significant correlates of self-reported PA. Conclusion： Measuring MVPA via self-report versus accelerometry produces considerably different results in a sample of young adolescents. Future studies should use caution when selecting outcome measures if the intent is to identify modifiable correlates of MVPA in youth.

A New Calculation of the In-Medium Quark Condensate at Finite Density and Temperature

The in-medium quark condensate is studied with an equivalent-mass approach in which one does not need to make assumptions on the derivatives of model parameters with respect to the quark current mass.It is shown that the condensate is generally a decreasing function of both the density and temperature with the decreasing speed depending on the confinement parameter.Specially,at given density,the condensate decreases on increasing temperature.The decreasing speed is comparatively small at lower temperature,and becomes very fast at higher temperature.

Variable Relation Parametric Model on Graphics Modelon for Collaboration Design

A new approach to variable relation parametric model for collaboration design based on the graphic modelon has been put forward. The paper gives a parametric description model of graphic modelon, and relating method for different graphic modelon based on variable constraint. At the same time, with the aim of engineering application in the collaboration design, the autonmous constraint in modelon and relative constraint between two modelons are given. Finally, with the tool of variable and relation dbase, the solving method of variable relating and variable-driven among different graphic modelon in a part, and double-acting variable relating parametric method among different parts for collaboration are given.

Yield Estimation Model of Citrus Based on Spectral Data and Agronomic Parameters

With the development of precision agriculture, the research that applies Remote Sensing technology, especially hyperspectral remote sensing, to realize crop management, monitoring and yield estimation, has been concerned. Nowadays, the growth-monitoring and yield-estimating methods in rice, wheat and other annual crops develop rapidly with some achievements having already been put into service. But the yield estimation research on perennial economic crops is few. Taking peren- nial citrus trees as the research object, using ASD spectrometer to collect citrus canopy spectral, this article studied and analyzed the citrus of veget＆tion index and its relationship on yield, synthetically considered the influence of the agriculture pa- rameters on crop yield, and finally constructed the citrus yield estimation model based on the spectral data and agronomic parameters. Through the Significance Test and Samples＇ Test, olutained that the model＇s fitting degree was R=0.631, F= 13.201, P〈0.01 and the error rate of estimating accuracy was controlled in the range 3%-16%, proving that the model has statistical signification and reliability. It concluded that hyperspectral acquired from citrus canopy has substantial potential for citrus yield estimation. This study is an application and exploration of Hyperspectral Remote Sensing technology in the citrus yield estimation.

Ventilation and climate simulation with the Multiflux code

Multiflux, a new thermal, hydrologic, and airflow model and software was being employed to solve the flow of heat, moisture, and air in and around an underground opening. The airway domain was solved with an integrated-parameter Computational Fluid Dynamic （CFD） module, which is an embedded part of the Multiflux code. The CFD model includes convection, conduction, and radiation for heat, as well as convection and diffusion for moist-ure transport in an air-filled opening. The surrounding rockmass model may be from any analytical solution, or from a complex thermal-hydrologic numerical model such as NUFT or TOUGH2. The rockmass model is interfaced to Multiflux using a novel technique called Numerical Transport Code Functionalization （NTCF）. The purpose was to briefly describe the Multiflux model and show four example applications. The first example reports the results of Multiflux simulations for a mine drift, comparing calculations with CLIMSIM, a well known mine climate software, and with measured data. The second and third examples involve development ends in two coal mines. Another development-end ventilation model in Multiflux is also shown as the 4th example compared with field measurements from the Lucky Friday Mine in Idaho, USA for comparison.The results of the study show very good agreement between the Multiflux model and the available measured field results.

Parametric modeling on spatial effect of excavation-damaged zone of underground cavern

Regarding excavation-damaged zone （EDZ） around underground opening as non-homogeneous rockmass with spatial deterioration effect on stuffiness and strength, a parametric model of EDZ using radius-displacement-dependent deformation modulus （RDDM） was proposed. Considering the nonlinearity characteristic of deformation and locality otherness of surrounding rock, deterioration parameter field of deformation modulus of rockmass around opening was quantitatively calculated through a given function. Applicability for multi-cavern condition and parameter sensibility of the model was analyzed by numerical experiments using synthetic data. Furthermore, the model was applied to identify EDZ of underground caverns of Pubugou hydropower station by calculating deterioration parameter field. Based on the parametric analysis of spatial effect and geological investigation, it is recognized that large radial deformation of deep fractured rock at the spandrel position and insufficient supporting bolts mainly result in great deformation pressure to act on the shotcrete and cause partial crack and spalling. It is shown that deterioration parameter field along the longitudinal axis of main powerhouse is evidently non-homogeneous in space and distributes exponentially along the radius from the opening. The model provides a simple and convenient way to identify the EDZ in the working state for rapid construction feedback analysis and support optimization of underground cavem from quantitative point of view and also aids in interpreting monitoring displacements and estimating support requirements.

Prediction of the Maximum Flow by the Model SASIS： Sensibility to the Empirical Parameters of the Shape of the Furrow

The model developed in this research presents effective mechanisms in simulations of a discharge strip understood between the minimum and the maximum allowable values, aiming to determine the relationship between discharge and water application efficiency, deep percolation and runoff rates, and consequently to optimize the performance of the furrow irrigation systems with continuous flow. The flow applied in each furrow must be adapted to the length, to the field slope and to the nature of the ground. The authors studied the maximum non erosive flow （Q,,,~）, in function of parameters obtained from the dimensions of the furrow, being Pl and/92, respectively, the linear and exponential parameters of the potential functions that described the relationship between the area of the cross section of flow （or wetted perimeter） and height of flow; in this way, the multiplicative effect of,01 on the area of the cross section of flow is linear, while ,02 is exponential. It verified a conjugated effect of,or and p20n the value of Q,,~. The results of this research point out for the importance of having an estimate of the parameters of the geometry of the cross section of flow （,01 and ,02） the most precise as possible, being known that the dimensions of this section can result in impracticable values of Qmax, outside of the acceptable strip in the literature, that is from 1.2 to 4.0 L.sl. This sensibility analysis was also of great benefit to create an interface in the software SASIS, capable to guide the user of this tool in the input of appropriate values for ,01 and P2 to the process of simulation of the irrigation for furrow with continuous flow and of the optimization of its performance.

New Li-ion Battery Evaluation Research Based on Thermal Property and Heat Generation Behavior of Battery

We do a new Li-ion battery evaluation research on the effects of cell resistance and polariza- tion on the energy loss in batteries based on thermal property and heat generation behavior of battery. Series of 18650 cells with different capacities and electrode materials are evalu- ated by measuring input and output energy which change with charge-discharge time and current. Based on the results of these tests, we build a model of energy loss in cells＇ charge- discharge process, which include Joule heat and polarization heat impact factors. It was reported that Joule heat was caused by cell resistance, which included De-resistance and reaction resistance, and reaction resistance could not be easily obtained through routine test method. Using this new method, we can get the total resistance R and the polarization parameter U. The relationship between R, η, and temperature is also investigated in order to build a general model for series of different Li-ion batteries, and the research can be used in the performance evaluation, state of charge prediction and the measuring of consistency of the batteries.

A back analysis of the temperature field in the combustion volume space during underground coal gasification

The exact shape and size of the gasification channel during underground coal gasification(UGC) are of vital importance for the safety and stability of the upper parts of the geological formation.In practice existing geological measurements are insufficient to obtain such information because the coal seam is typically deeply buried and the geological conditions are often complex.This paper introduces a cylindrical model for the gasification channel.The rock and soil masses are assumed to be homogeneous and isotropic and the effect of seepage on the temperature field was neglected.The theory of heat conduction was used to write the equation predicting the temperature field around the gasification channel.The idea of an excess temperature was introduced to solve the equations.Applying this model to UCG in the field for an influence radius,r,of 70 m gave the model parameters,u1,2,3...,of 2.4,5.5,8.7...By adjusting the radius(2,4,or 6 m) reasonable temperatures of the gasification channel were found for 4 m.The temperature distribution in the vertical direction,and the combustion volume,were also calculated.Comparison to field measurements shows that the results obtained from the proposed model are very close to practice.

Remote sensing Extraction model of redtide biomass by airborne hyperspectral technique

Our work is based on the known research results of inherent optical quality of ocean color constituents.According to optimized parameters and induced fluorescence term of chlorophyll, this paper puts forward a remote sensing reflectance model of sea water, which is fitted in Liaodong Bay of Bohai. An inverse model that can evaluate redtide biomass according to chlorophyll retrieval is provided by inducing a functional extreme problem. The calculation example of the model indicates that the inversion model has explicit mathematic and physical meaning, but its practicability needs to be verified.

Frame complexity optimized selection for H.264/AVC video coding

To improve the coding performance of H.264/AVC, this paper proposes a rate control scheme composed of a novel flame complexity optimized selection and a quantization parameter （QP） value computation approach. First, it extracts the frame coding complexity from two rate distortion models, and then introduces five statistic modes to estimate the frame coding complexity. An optimal mode is selected according to the coding efficiency. Finally the paper presents a novel QP calculation method for the H.264/AVC rate control. Experimental results show that the proposed algorithra outperforms the algorithm integrated in the 3M model in obtaining precise frame coding complexity, achieving robust buffer control and improving coding quality. And the improving visual quality is high up to 0.90dB for CIF sequences.

Online Monitoring the Products Quality by Measuring Cavity Pressure during Injection Molding

Injection molding is a complicated production technique for the manufacturing of polymer products. During injection molding, it＇s hard to predict molding quality; the injection molding parameters, such as mold temperature, melt temperature, packing pressure and packing time, affect the final properties of product. The cavity pressure is a significant key factor. Residual stress and injection molding weight are significantly affected by the cavity pressure. This study created an approach to predict weight of injection-molded by real-time online cavity pressure monitoring. This study uses a 6-inch with thickness lmm light guide panel and the largest area beneath the pressure curve of time as well as the maximum pressure as its characteristic. The upper and lower limits of the control are set to ＋2 standard deviations, and GUI （Graphical User Interface）-based LabVIEW software is used to perform calculation and analysis of the pressure curve. The results of the experiment show that the online internal cavity pressure monitoring system can effectively monitor the quality of the molded products. In 500 injection molding cycle tests, its error rate was less than 8%, whereas the deviation in mass of the molded products selected through the system＇s filtering process was successfully controlled to be within ±4%.