49 results about "Model inversion" patented technology

The invention provides a ground-air integrated agricultural monitoring system and a ground-air integrated agricultural monitoring method, wherein the system comprises an aerial agricultural monitoring device, at least one ground agricultural monitoring device and a terminal server. The aerial agricultural monitoring device is used for obtaining crop canopy data; the least one ground agricultural monitoring device is used for obtaining crop growing environment data and crop physiological and biochemical data; and the terminal server is used for fusing the crop canopy data, the crop growing environment data and the crop physiological and biochemical data of the same monitoring point obtained by the aerial agricultural monitoring device and the ground agricultural monitoring device, and performing crop growth model inversion; moreover, the aerial agricultural monitoring device and the ground agricultural monitoring device are respectively in wireless communication with the terminal server. According to the invention, physiological and biochemical parameters and growing environment of the crops are dynamically monitored in real time in both the ground and the air at the same time, and the monitoring data is fused, and thus, multi-parameter monitoring and large-area efficient monitoring on the crops are implemented.

The invention discloses a global gravity field model inversion method. The method comprises the following steps: step 1: acquiring low-orbit gravity satellite data, and preprocessing the low-orbit gravity satellite data; step 2: calculating a pure Earth gravitation acceleration observation value by virtue of numerical differentiation; step 3: establishing a novel observation equation according toan observation value constant weight strategy; and step 4: estimating parameters, and obtaining an Earth gravity field model position coefficient. According to the acceleration domain residual error analysis of a reference model, a symmetric Toeplitz covariance matrix is established by directly adopting an autocorrelation sequence as a basic element, a complicated operation process of a high-dimensional huge matrix can be effectively avoided when the acceleration covariance matrix is calculated by an orbit prior dense array by virtue of an error propagation law, and the time for fixing the weight can be reduced, so that the calculation efficiency of the gravity field model can be improved. Compared with a classical method, an algorithm provided by the invention is simple and efficient in operation; and moreover, the stability and reliability of the inversion calculation efficiency of the gravity field model can be ensured.

The invention, which belongs to the technical field of underground medium speed model inversion of seismic exploration, relates to a full-waveform inversion method based on a seismic record integral. The method comprises: an observed seismic record is obtained, denoising processing is carried out, and a seismic wavelet is extracted; a simulated seismic record is calculated; time-domain integrating processing is carried out on the observed seismic record after denoising and the simulated seismic record; difference processing is carried out two groups of data after integrating to form a back transmission source, and a back transmission wave field; a forward transmission wave field is calculated; a gradient is calculated; a conjugate gradient is calculated; according to a linear searching method, a step length of iterative updating is calculated; a model correction value is calculated and a speed model is updated; and whether the updated model meets a given condition is determined; if so, a result is outputted; and if not, the updated speed model is used as a new initial speed model and calculation is started again. Therefore, a problem that the inversion is trapped into a local limit value problem because of lack of low frequency seismic data and thus inversion fails during full waveform inversion is solved.

The invention relates to a landslide monitoring and early identification method based on multi-data-source SBAS technology. The method comprises the steps of obtaining multiple SAR image data coveringthe same time period of a monitoring range; common main image selecting; SAR image registration; selecting of interference pairs; generating of a differential interferogram; ground point selecting; estimating of residual topographical phases; removing atmospheric phases and residual errors; and establishing a model-inversion sedimentation rate. In the invention, according to the geographic locations and topographical features, existing different source SAR data of the same region is utilized to carry out landslide deformation monitoring based on small baseline subset technology, and meanwhileground point selection constraints are increased to strengthen the connection between different source data. Differences between the results are narrowed, and the high-precision and high-accuracy inversion results without other checking conditions are achieved, thus providing effective and reliable technical support for the discovery and prevention of landslide geological disasters.

The invention discloses a technical method for dynamic monitoring of carbon storage in forest ecosystems. The steps are: firstly, obtain the forest ecosystem biomass according to the field survey data, and then calculate the forest ecosystem carbon storage from the forest ecosystem biomass; secondly, use the forest ecosystem carbon storage and remote sensing The relationship between the gray value of the image and the factors of the geographic information system, the carbon storage inversion model during the field survey period was established, and the carbon storage of the forest ecosystem during the field survey period was obtained from the inversion; finally, the image during the field survey period was used as a reference image , carry out relative atmospheric correction on regional historical remote sensing images and reference images, and use the carbon stock inversion model during the field survey period to calculate the carbon stock of the forest ecosystem in each historical period of the region, so as to achieve accurate , Rapid dynamic monitoring of forest ecosystem carbon storage.

The invention relates to a pavement deflection measuring system and a pavement deflection measuring method. The system comprises speed sensors, angular speed sensors, an acceleration meter, a mileage encoder and a processor, wherein a plurality of speed sensors are respectively arranged on load cross beams in different distance away from a load origin and used for measuring the changed speeds of the load cross beams where the speed sensors are positioned in the longitudinal direction; the angular speed sensors are used for measuring the rotating amplitude variations of the load cross beams; the acceleration meter is used for measuring the pitch acceleration value of a measuring vehicle; and the mileage encoder is used for recording the running mileage information of the deflection measuring vehicle. The measuring method comprises the following steps of: measuring the pavement deflection speeds of multiple points of a pavement by adopting the plurality of co-beam speed sensors, and calculating the dynamic deflection value of the pavement according to the elastic deformation theory of the pavement and model inversion. The system and the method are suitable for continuous sampling and measuring under dynamic conditions, are used for quickly and efficiently measuring the deflection value, saves manpower and material resources, and are more accurate and reliable in comprehensive estimation of road sections.

The invention provides a poststack acoustic wave impedance inversion method and belongs to the oil gas and coal bed gas earthquake exploration and development field. The method comprises the following steps: (1) processing poststack earthquake data so that a frequency spectrum of the poststack earthquake data is matched with a frequency spectrum of an acoustic wave impedance acquired from log data and relative acoustic wave impedance data is obtained; (2) constructing an initial acoustic wave impedance model based on model inversion or a variation function; (3) carrying out forward modeling calculation on the initial acoustic wave impedance model so as to acquire synthetic earthquake data; (4) comparing the synthetic earthquake data and actual poststack earthquake data so as to calculate modification of the acoustic wave impedance model, using the modification to modify and update the initial acoustic wave impedance model so as to acquire an updated acoustic wave impedance model, then carrying out the forward modeling calculation on the updated acoustic wave impedance model so as to acquire synthetic earthquake data, taking the model as the initial acoustic wave impedance model, and repeating the step 4 so as to acquire a final inversion result.

The embodiment of the invention provides a method and device for determining the three-dimensional stratum pore pressure of a target area. The method comprises the steps that 1, the first incidence relation corresponding to the normal compaction trend speed is determined by combining the Philip algorithm with the Eaton algorithm; 2, a three-dimensional speed body and a three-dimensional density body are obtained on the basis of low-frequency initial model inversion of the target area; 3, the longitudinal wave speed, the stratum speed, the overlying stratum pressure and the hydrostatic pressure of the target area are determined according to the three-dimensional speed body and the three-dimensional density body; 4, the normal compaction trend speed of the target area is worked out on the basis of the first incidence relation as well as the longitudinal wave speed, the stratum speed, the overlying stratum pressure and the hydrostatic pressure of the target area; 5, the Eaton index is determined through the single well drilling measured data of the target area; and 6, the three-dimensional stratum pore pressure of the target area is determined on the basis of the Eaton algorithm and the normal compaction trend speed of the target area. According to the technical scheme provided by the embodiment of the invention, the prediction precision of the stratum pore pressure can be improved, and data support is provided for oil-gas exploration work and data interpretation work.

A system and method for generating test cases includes a model inverter configured to generate an inverse of a model, a conversion module configured to convert a requirements specification into at least one output scenario and a model simulator configured to generate a test case that satisfies the requirements specification based on the inverse of the model and the at least one output scenario.

The invention discloses a multi-scale thermal inertia model inversion-based method for measuring the moisture content of surface soil. The method for measuring the moisture content of the surface soilcomprises the following steps: performing discrete wavelet transform analysis on an acquired ASTER thermal infrared image to acquire a high-definition ASTER thermal infrared image; acquiring and calculating a surface component temperature according to the high-definition ASTER thermal infrared image; calculating a surface albedo through a weight function for converting a narrow band albedo into awide band albedo; estimating a surface sensible heat flux according to the high-definition ATER thermal infrared image; calculating a surface latent heat flux by an energy residual method according to the surface sensible heat flux; calculating a thermal inertia of the soil according to the surface component temperature, the surface sensible heat flux, the surface latent heat flux and the surfacealbedo; and building a surface soil moisture content inversion model according to the thermal inertia of the soil. By implementation of the method of the embodiment, the measurement precision of themoisture content of the soil within a detected region can be effectively improved.

The application discloses an EOS-ELM-based piezoelectric driving positioning platform modeling method and system, and an EOS-ELM-based piezoelectric driving positioning platform control method. According to the application, by taking the current desired output displacement, a plurality of previous output displacements and an input driving voltage as the input values of a model, and taking the input driving voltage value of a piezoelectric driver at the current moment as an output value of the model, a hysteretic nonlinear multi-value mapping problem is solved, and a complicated model inversion process can be avoided; secondly, the weights and the threshold values between a hidden layer and an input layer are given randomly, and a conventional neural network can be transformed into a linear equation model, so that the training time of the model is shortened greatly; thirdly, an infinite differentiable function is used as a hidden layer activation function, so that an excellent training effect can be realized; additionally, just by adopting a recursive least square method, the parameter online adaptive updating can be realized, and the applicability of the model is improved; furthermore, by integrating a plurality of weak learning machines, the precision and the stability of the model can be improved.

The invention discloses an OS-ELM based modeling and control method and system of a piezoelectric ceramic driver. According to the invention, currently expected output displacement, previous multiple output displacement and input driving voltage are used as model input values, and the input driving voltage values at the present are used as model output values, so a hysteresis nonlinearity multi-valued mapping problem is solved, and the output values can be directly used for driving the piezoelectric ceramic driver so that tedious model inversion solving processes are avoided; then, a weight and a threshold value between a hidden layer and an input layer are randomly given, the model is converted into a linear model, and a connection weight between the hidden layer and the output layer can be further calculated only by adopting the simple generalized inverse calculation, so training time is greatly shortened; an infinite differentiable function is used as an activation function of the hidden layer, so quite high precision is achieved; and furthermore, online adaptive updating of parameters can be achieved only by using the recursive least squares method, thereby facilitating improving of applicable performance of the model.

The invention discloses a voice classification method through using vocal cord modeling inversion. All kinds of voices are effectively distinguished from the perspective of a sound production mechanism. The method is characterized by mainly using complex cepstrum phase decomposition to obtain an actual voice glottis wave as a target glottis wave; carrying out vocal cord dynamic model inversion operation through a matching target and a model glottis wave characteristic parameter through using an optimization algorithm; and selecting a normal voice and a special voice to carry out identificationand classification so as to achieve better accuracy. In the invention, after an actual voice signal is input, the actual glottis wave is extracted and is taken as a target, and a genetic algorithm isused to carry out inversion and optimize an original model so as to simulate vocal cord vibration conditions during the sound production of different voices. An experiment result shows that the matched relative error of each characteristic parameter does not exceed 1.95% after model inversion and an inversion effect is good. The normal voice and the special voice are selected to carry out identification and analysis and better accuracy can be achieved.

The invention discloses a spacecraft transient thermal analysis model inversion correction method. The spacecraft transient thermal analysis model inversion correction method comprises the main steps of firstly establishing a spacecraft transient thermal analysis model and counting to-be-corrected heat transfer parameter sets of the model and an uncertainty range of the model; then carrying out random sampling and establishing as well as optimizing original samples; building a transient temperature analysis value and experiment value error objective function, and transforming the function into a fitness function, and correcting the parameter by adopting a loop inversion method which combines a genetic algorithm and a quasi-linear algorithm (BFGS); and finally selecting multiple groups of better results, and carrying out positive optimization by utilizing experimental data unused in the correction so as to obtain a final heat transfer parameter correction value. The invention provides the loop inversion method which combines a universal objective function construction method, the genetic algorithm and the quasi-linear algorithm (BFGS) of the correction method, and a positive optimization method for an approximate optimal result; and the correction precision and correction efficiency of the spacecraft transient thermal analysis model parameter are improved while an inversion false solution is avoided to the greatest extent.

The invention relates to a spatial relation modeling method for remote sensing image objects. The method comprises the steps of S1 image segmentation, S2 training sample selection, S3 initial classification result generating, S4 spatial modeling image sample object selection, S5 calculating of the spatial distance between image sample objects, S6 covariance function estimation of image sample objects, S7 point support model inversion, S8 establishing of spatial relationship between any two image objects, and S9 implementation of object-oriented geostatistical weighted KNN classification method. The method provided by the invention has the advantages that the spatial relation modeling method for the remote sensing image objects is proposed for the first time to fill the gap in the field; compared with a modeling method which does not consider the data characteristics of the image objects, the method effectively improves the classification accuracy; and the method is not limited to the application of the field of remote sensing image classification, and can be widely applied to the modeling of a variety of image objects.

The invention discloses a gas pipeline leakage point positioning method based on acoustic frequency shift inversion. The method comprises the following steps that direct wave signals generated in a gas pipeline leakage point is received, arrayed and detected, then the direct wave signals received by each wave detector is subjected to spectrum analysis, and the centroid frequency of each direct wave signal is calculated to obtain a centroid frequency-distance fit curve (xi,); and according to the obtained centroid frequency-distance fit curve, the transverse position of the gas pipeline leakagepoint is determined through observing the change rule of the curve, and a genetic algorithm is adopted to carry out model inversion, wherein the algorithm adopts genetic operation with variation, crossing and selection, so that superior data are selected, inferior ones are eliminated, and repeated iteration is carried out, and finally the optimal solution is found until the depth position of thegas pipeline leakage point is determined. According to the gas pipeline leakage point positioning method based on acoustic frequency shift inversion, the attenuation characteristics of the gas leakagesound wave centroid frequencies is utilized for leakage point positioning, so that the positioning precision of the leakage point is effectively improved, and meanwhile the anti-interference capability is high.

The invention discloses a process method for rapidly developing a cast product sample, belongs to the technical field of casting. The technical scheme in the invention is as follows: 1. using a plurality of general devices to manufacture resin sand or coated sand billet similar with cast product specification; 2. making a model inversion on a model sand billet, processing a core sand billet as a required sand core; 3. for the cast product sample with complex structure and shape, using a method for processing the cast product sample with complex structure and shape as the model inversion and the sand core on a numerical control machine such as a machining centre or a carved milling machine; 4. coating the processed model inversion and sand core with bad surface quality by using alcohol-based or water-based paint until meeting the demand; 5. making a pouring and feeding system, combining the model inversion and sand core, pouring to obtain a corresponding cast product sample. The process method effectively shortens the manufacturing period of the product sample and core box, reduces the product development input as much as possible at an early stage, and lays foundation for manufacturing a permanent metal model and core box in the next step.

The invention relates to a reservoir prediction method and device based on a synthetic seismogram, and belongs to the technical field of petroleum exploration development. A density-speed relationshipempirical formula suitable for a research area is searched to carry out the environment correction of a mudstone collapse section on well logging data in the research area to obtain a corrected density well logging curve and acoustic logging curve, according to the corrected density well logging curve, a synthetic seismogram record is manufactured to obtain a synthetic record which is more matched with a well byway, so that important basic data is provided for lowering multiplicity and improving inversion accuracy for subsequent model inversion, and therefore, reservoir predication accuracy is further improved.