27 results about "Non linearity" patented technology

The invention discloses an inverted coplanarity detection method based on 3D vision and refraction compensation, and the method comprises the following steps: 1, selecting an optical measuring head of which the measuring range can cover the whole bottom surface of a measured object according to the size of the bottom surface of the measured object, installing the optical measuring head below glass, and enabling light to penetrate through the glass from bottom to top; 2, establishing a glass refraction compensation model; 3, correcting the glass refraction compensation model; 4, collecting three-dimensional point cloud data of the bottom surface of the measured object; 5, point cloud extraction and processing; step 6, calculating and evaluating the coplanarity; and 7, repeating the steps 4-6 to realize coplanarity analysis under different conditions. Through fusion of a theoretical model and experimental calibration, nonlinear error measurement caused by factors such as glass refraction and non-ideal installation is solved, the measurement precision is improved, the reference plane is established by adopting a minimum area method, the coplanarity evaluation reasonability is improved, and the method can be expanded and applied to all occasions needing to measure the coplanarity of the bottom structure from bottom to top.

The invention discloses an automatic discretization method for a hydroelectric nonlinear characteristic curve. The method comprises the following steps: collecting n sample points of a hydroelectric original nonlinear characteristic curve, and setting the number of required discrete points to be m; virtually connecting the head and tail points of the original nonlinear characteristic curve to form a straight line, calculating the vertical distances between all sample points on the original nonlinear characteristic curve and the straight line, and calculating the maximum distance value D0 corresponding to the maximum distance point P0; if the maximum distance value D0 is larger than the distance threshold value, the maximum distance point P0 is reserved; respectively screening out the maximum distance points in the residual sample points, and comparing the maximum distance points with a distance threshold value; if the maximum distance value D0 is smaller than or equal to the distance threshold value, all middle points are removed, and only the starting point and the ending point are reserved; and repeating the steps, and removing the point with the minimum error every time until only m discrete points remain. According to the method, the discrete points with the given number can be rapidly and automatically extracted, and the extracted discrete points can accurately represent the original nonlinear characteristic curve.

The invention provides a load demand prediction method and apparatus, an electronic device and a storage medium. The method comprises the steps of obtaining first load data; performing time sequence decomposition on the first load data to obtain seasonal data, trend data and residual data; carrying out prediction sample extraction based on seasonal data, trend data and residual data; inputting the prediction sample into a first model to obtain first prediction data; performing Pearson chi-square test calculation according to the prediction sample and the first prediction data to obtain a significance level value; when the significance level is greater than a first preset threshold value, determining that the final output is first prediction data; and when the significance level is not greater than a first preset threshold value, inputting the prediction sample into a second model to obtain second prediction data, and determining the final output as the second prediction data. Therefore, non-linear trends such as seasonality and residual errors are considered during subsequent load demand prediction.

The invention discloses a magnetic tunnel junction and a TMR sensor, and belongs to the technical field of magnetic sensing. The magnetic tunnel junction comprises a substrate layer, a ferromagnetic free layer, a barrier layer, a ferromagnetic reference layer and an antiferromagnetic pinning layer, wherein the ferromagnetic free layer, the barrier layer, the ferromagnetic reference layer and the antiferromagnetic pinning layer are sequentially stacked on one side of the substrate layer; wherein the crystal face of the ferromagnetic free layer is a (110) crystal face. According to the embodiment of the invention, the linearity of the magnetic field sensing of the magnetic tunnel junction can be improved, the nonlinear error is reduced, and the TMR sensor can obtain a wider high-linearity measurement range.

This invention relates to the field of earthwork trenching technology, specifically to a method and system for excavating slope trenches for prefabricated quick-assembly grid beams. In this invention, by collecting high-density discrete coordinates of the slope and constructing a reference straight-line vector connecting the beginning and end, the geometric correction depth of each surface point relative to the ideal flat bottom surface is calculated. This eliminates nonlinear errors caused by the original slope undulations, ensuring that the bottom surface of the trench accurately adapts to the straight-line configuration of the prefabricated components. The resistance gradient change characteristics during the excavation process are combined with a clustering algorithm to dynamically lock the interface between soft and hard geological layers. The excavation depth required by the geometric design is fused and compared with the measured depth of the geotechnical boundary in the same coordinate system. Based on the comparison results, differentiated cutting or stripping commands are intelligently generated. When encountering hard rock layers higher than the designed bottom surface, a fuzzy PID algorithm is activated to enhance hydraulic driving force and ensure cutting efficiency. In soft soil operating areas, standard pressure is maintained to reduce energy consumption.

A Gaussian transfer function is applied to a profile of a feature on an image of a workpiece generated using an electron beam metrology tool. This generates an output profile. Non-linearity of the output profile is determined. Two threshold values can be selected and plotted on an XY plane. Proximity of a slope of the threshold values to a comparison value is determined.

The invention provides a Mars landing section autonomous navigation method and device fusing point and line features. The method comprises the following steps: acquiring measurement information of the surface of Mars; analyzing point features and line features in the measurement information, and constructing a measurement model; constructing a state model according to the measurement information and a dynamical model; establishing an autonomous navigation model by using the measurement model and the state model; and updating the navigation state through extended Kalman filtering according to the autonomous navigation model. The advantages and disadvantages of the point features and the line features in different scenes are analyzed, the fact that the line features have high robustness compared with the point features in the low-texture and complex environment is analyzed, and a nonlinear error with higher precision is introduced to define the measurement error of the line features. A digital simulation result shows that compared with a traditional measurement model which only uses point feature measurement, the method has higher navigation precision.

The present invention discloses an ultra-precision global bridge rotation angle measurement method, which relates to the technical field of angle measurement. The method includes: continuously capturing a sequence of bridge motion images and recording timestamps. Selecting key points on the bridge deck through a feature detection algorithm, constructing a spatio-temporal data matrix and performing data normalization processing. Extracting the global features of the overall motion of the bridge deck by using linear principal component analysis; extracting the local non-linear motion features in the residuals by using non-linear principal component analysis. Calculating the displacement of the feature points based on the reconstruction matrix in combination with the LK optical flow method. Fitting the displacement data by using the least squares method, and calculating the angle based on the arctangent function on this basis. This method has the characteristics of high precision and global coverage, is applicable to fields such as bridge health monitoring and safety assessment, and can effectively capture the overall and local motion characteristics of the bridge.

The present invention relates to a method for registering large tilt aerial images and orthophotos. The technical solution is as follows: First, use the ASIFT algorithm to simulate the affine image perspective of the large tilt aerial image S to obtain the simulated affine image S'; then construct a non-linear scale space for the simulated affine image S' and the orthophoto R; in the non-linear scale space, use phase consistency and the Sobel operator to construct a feature point metric matrix for feature point extraction, and construct a GLOH-like feature description operator by calculating the amplitude and direction of the feature points; then perform feature registration and error elimination on the simulated affine image S' and the orthophoto R to obtain correctly registered corresponding points. The corresponding points in the large tilt aerial image S are obtained from the correctly registered corresponding points of the affine simulated image S' through the simulation parameters, and finally, the large tilt aerial image S and the orthophoto R are corrected according to the affine transformation parameters Ψ. The present invention has the characteristics of small registration error, good anti-noise performance, high success rate and stable results.

The purpose of the present invention is to improve drawing accuracy by accurately adjusting the deflection sensitivity of a deflector.　This method for adjusting the deflection position of a charged particle beam, in a charged particle beam drawing device, includes: a step (S102) for measuring, at a predetermined pitch using a laser interferometer, the position of a stage on which a substrate to draw on is placed; a step (S105) for scanning, with the charged particle beam, a mark on the stage while moving the stage by a predetermined amount, using a stage position measurement result obtained by the laser interferometer; a step (S106) for detecting the position of the mark; a step (S108) for acquiring non-linear error information of the position of the mark that depends on the laser interferometer; a step for acquiring a plurality of positional deviation amounts on the basis of the measured position of the stage and the detected position of the mark using the non-linear error information; and a step (S109) for adjusting the deflection position of the charged particle beam on the basis of the acquired plurality of positional deviation amounts.

The invention discloses a tail water taking elevation identification method. The method comprises the following steps: obtaining dam front water level data, reservoir vertical water temperature distribution observation data and tail water temperature data; cleaning the water temperature distribution observation data, and removing observation points higher than the water level in front of the dam and invalid data; dividing the cleaned water temperature distribution observation data into a plurality of layers; respectively solving elevation fitting functions of the multiple layers and sections, and splicing the elevation fitting functions to form an elevation high-precision fitting function of the continuous temperature; according to the elevation high-precision fitting function, constructing an inversion objective function, and calculating the water taking elevation of the tail water under the dam; according to the method, the thermal stratification structure is distinguished, the continuous high-precision temperature elevation function is constructed, the error objective function is constructed in combination with the temperature of the tail water, then the water taking elevation is inverted through the numerical optimization method, nonlinear characteristics such as a thermal spring layer are effectively captured, and the accuracy of water taking elevation inversion is remarkably improved.

The present invention discloses a gas detection method, which relates to the field of gas detection. The initial concentrations of carbon dioxide and oxygen at the current moment and the denoised air pressure processed by the Kalman filtering algorithm are measured and obtained. A model-driven backpropagation network is used to construct a correction model for air pressure based on the Beer-Lambert law and the Nernst equation. The denoised air pressure at the current moment is substituted into the correction model, and a first corrected representation is jointly generated in cooperation with the non-linear mapping of the initial concentration at the current moment. A second hidden representation is generated based on the non-linear correlation mapping of the first corrected representation, the measurement device parameters, and the denoised air pressure at the current moment. The concentration at the current moment is obtained through linear modulation. Through the cooperative pressure compensation combining physical model driving and data driving, the pressure correction accuracy is improved, and the accurate detection of carbon dioxide and oxygen in a high-pressure environment is realized.

A high-robustness image stitching method disclosed by the present invention includes: dividing a preprocessed sequence of image sets into a reference image and an image to be stitched; constructing a non-linear scale space for feature detection; using a multi-scale depth hybrid feature descriptor HMD to describe feature points; obtaining an initial matching set and a candidate matching set; using a guided maximum likelihood random consensus algorithm to purify the candidate matching set and obtaining optimal model parameters through iterative calculation; eliminating the matching set that does not conform to the best parameter model; aligning the images through perspective transformation and finding the position of the best stitching line; determining the best stitching line and obtaining a seamless stitched panoramic image based on multi-band weighted fusion; outputting the final result after cropping the edges of the panoramic image. The present invention integrates the multi-scale and depth information of the target scene, effectively improves the stitching quality of panoramic images and the accuracy of geometric correction, suppresses the artifact defects caused by repeated textures and occlusions, and has higher stitching accuracy and robustness.

The invention relates to the technical field of soil detection, and discloses a soil heavy metal pollution detection method and system based on spectral data, and the detection method comprises the steps: obtaining a spectral reflectivity image of a soil surface layer of a to-be-detected region, and extracting a hyperspectral feature sequence to construct a spectral response curve of target soil; calculating a nonlinear mapping characteristic factor group and a credibility factor group based on the spectral response curve; constructing a credibility representation value according to the two, and dividing a reliability grade label; and judging whether model fitting distortion exists, whether a local fitting sub-model is called and whether model weight adjustment is carried out according to the label, and outputting a final prediction concentration result. The method disclosed by the invention has relatively high detection accuracy on relatively low-content soil heavy metals.

A Gaussian transfer function is applied to a profile of a feature on an image of a workpiece generated using an electron beam metrology tool. This generates an output profile. Non-linearity of the output profile is determined. Two threshold values can be selected and plotted on an XY plane. Proximity of a slope of the threshold values to a comparison value is determined.

The invention relates to the technical field of geometric quantity precision measurement, and discloses a four-layer half-hole plate precision drilling and cutting track measurement system, which comprises an image acquisition module, a deformation analysis unit and a track measurement module, and is characterized in that the image acquisition module acquires a random texture image of the surface of a to-be-drilled site and a dynamic projection contour of a cutting tool; the deformation analysis unit identifies a feature point displacement vector and is used for reconstructing a non-uniform grid and converting the deformation of the base material into a local coordinate offset of a corrected reference surface; and the trajectory metering module extracts gray gradient data to determine a main shaft deflection vector, and identifies a depth zero position in combination with a shadow shortening rate, and by using an inversion mechanism of base material local texture displacement and rotating light and shadow characteristics, the non-linear residual caused by press fit stress release is eliminated, and the spatial consistency of complex inner layer space motion trajectory metering is ensured.

The invention belongs to the technical field of radiation measurement, and particularly relates to a neutron gamma pulse rapid matching method and system. The method comprises the following steps: acquiring a digital pulse sequence; determining a parabolic virtual baseline according to the weighted sum of the pulse baseline noise and the peak amplitude, and solving an intersection point of the pulse first-order differential signal and the baseline as a weighted zero crossing point time feature; calculating a ratio of waveform area first-order moments before and after a peak point as an asymmetric torque characteristic; constructing a basic transition probability matrix by using a historical sample, and performing nonlinear disturbance according to the asymmetric torque characteristic of the current pulse to generate an instantaneous transition probability matrix; and combining the two features to form a two-dimensional vector, calculating a weighted mahalanobis distance based on an instantaneous transition probability matrix, and completing matching according to a short distance principle. According to the invention, the problem of insufficient precision and stability caused by noise sensitivity and fixed matching criterion of the existing pulse discrimination technology is solved.

The controller (41) is configured to emit a command to the light string (1) to present a light effect. The command specifies one or more colors and / or light intensities. The controller or the light string is configured to receive a first input signal specifying a first spatial distribution. The first spatial distribution indicates a first position of a light source (11-19) of the string relative to the object. The light string is configured to: receive a command; in response to the command, determining a first mapping of light settings to the light source based on the one or more colors and / or light intensities and also based on the first spatial distribution; and controlling the light source to present the light setting according to the first mapping. The first mapping corresponds to a non-linear or otherwise unequal light setting distribution on the light source.

The invention discloses a saddle opening lineation solution, a tool is manufactured according to the combination of the principle that an approximation method approaches to a circle and the principle of a projection method, the tool comprises a positioning groove and a probe, a shape can be accurately drawn on a nonlinear plane, and the innovation is that the concept that the approximation method approaches to the circle is adopted, a line segment is used for replacing the drawing of the circle, and the drawing accuracy is high. The approximation method combines an approximation circle and a projection method, the approximation method makes up the limitation that the projection method can only draw a line on a preceding plane, and a standard shape contour can be drawn on a non-linear plane. The multipurpose saddle opening marking tool is characterized in that a tool is manufactured according to the combination of the principle that an approximation method approaches a circle and the principle of a projection method, the tool comprises a positioning groove and a probe, and a shape is accurately marked on a nonlinear plane.

The present invention discloses a method for resolving multiple echoes in the air of a lidar based on distance and reflectivity constraints. Geometric calibration equations and radiation calibration equations of a mapping lidar are respectively established to correct the distance measurement indication value, angle measurement indication value, amplitude, and relative reflectivity of each echo. The time-of-flight principle is used to determine the time interval of the farthest laser echo. According to the laser emission repetition frequency, the number of main laser emission waves existing within the time interval of the farthest laser echo is determined. Then, adjacent consecutive scanning points in the pitch and azimuth directions are respectively determined according to the scanning pitch angle and azimuth angle. Finally, the continuity of the main emission wave is judged, and based on the geometric and radiation continuity between adjacent consecutive scanning points, the continuity noise energy is calculated using the ranging result and relative reflectivity result of each echo, and the continuity noise function corresponding to each MTA is obtained. A robust non-linear optimization model is constructed and optimized to solve. The present invention can achieve accurate calculation of the MTA time.

The invention discloses a discontinuous distortion correction method based on data driving, and the method comprises the steps: collecting uniformly-distributed training data in a view field range through employing a precise angle measurement method, and enabling the training data to comprise image point distortion coordinates of a multi-detector splicing imaging system and corresponding ideal coordinates; constructing a multi-branch deep learning network model, wherein the multi-branch deep learning network model comprises a plurality of special branches, a fusion layer, a sharing layer and an output layer; preprocessing training data obtained by precise angle measurement, and training the multi-branch deep learning network model by using the preprocessed training data; and using the trained multi-branch deep learning network model to obtain the corrected coordinates of any input actual measurement coordinate point. According to the invention, the problem of difficulty in accurate representation of holographic distortion caused by optical distortion discontinuity and nonlinearity in a current multi-lens multi-detector splicing imaging system can be solved.

Processing circuitry receives coded information of a current block indicative of applying LIC on the current block according to at least a first reference block in a first reference picture. The processing circuitry determines, for a sample in the current block, at least a first reference sample in the first reference block, the first reference sample is co-located to the sample in the current block. The processing circuitry calculates a weighted sum of a plurality of terms and an offset for the LIC according to a plurality of parameter values for a plurality of parameters used in the LIC, and reconstructs the sample accordingly. The plurality of parameter values includes at least a first weighting value for a first weighting that is applied to a non-linear term of the first reference sample of a kth power, k is a power index value and is not equal to 1.

The present application belongs to the technical field of radiation measurement, and particularly relates to a neutron gamma pulse fast matching method and system. The method comprises: obtaining a digital pulse sequence; determining a parabolic virtual baseline according to the weighted sum of the pulse baseline noise and peak amplitude, and solving the intersection point of the first-order differential signal of the pulse and the baseline as the weighted zero-crossing time feature; calculating the ratio of the first moments of the waveform areas before and after the peak point as the asymmetric torque feature; constructing a basic transition probability matrix using historical samples, and generating an instantaneous transition probability matrix according to the nonlinear disturbance of the asymmetric torque feature of the current pulse; combining the two features to form a two-dimensional vector, calculating the weighted Mahalanobis distance based on the instantaneous transition probability matrix, and completing the matching according to the principle of close distance. The present application solves the problems of noise sensitivity, fixed matching criteria, insufficient precision and stability of the existing pulse discrimination technology.

The invention relates to the technical field of plate measurement, and relates to a plate size measurement method and system based on three-dimensional modeling, and the method comprises the steps: determining at least three non-linear reference points on a roller of a roller path, and obtaining a three-dimensional model of the roller; obtaining a first coordinate of a nonlinear reference point measured by a laser scanner; obtaining a corresponding second coordinate of a nonlinear reference point in the three-dimensional model; calculating a coordinate transformation matrix according to the first coordinate and the second coordinate of the nonlinear reference point; acquiring a plurality of acquisition coordinate points on the upper surface of the roller of the roller path acquired by the laser scanner, converting the acquisition coordinate points into coordinate points in a model coordinate system through a coordinate transformation matrix, and recording the coordinate points as model coordinate points; screening model coordinate points whose distances to the three-dimensional model are greater than a preset distance threshold, and recording the model coordinate points as target model coordinate points; and calculating the size of the plate. According to the invention, the size of the plate can be accurately measured in a poor illumination environment.

A method and system for dynamically representing the separation for air traffic has been developed. First, air traffic is detected which requires maintenance of a separation distance from an ownship aircraft. The ground speed of the ownship, the ground speed of the air traffic and the current separation distance is determined. A predicted separation distance is calculated following a specific time interval. The predicted separation distance between the air traffic and the ownship is based on a differential in ground speed between the air traffic and the ownship and the specific time interval. The location of the air traffic, the separation distance and the predicted separation distance are all shown on a graphical display onboard the ownship. The separation distance and the predicted separation distance are represented on a non-linear scale on the graphical display.

This application relates to a code density calibration method and system based on dark counting. The method includes: setting a SPAD in Geiger mode under no-light conditions; generating a Poisson-distributed dark counting pulse unrelated to the TDC clock using thermal perturbation; using this pulse as a trigger signal to acquire and convert raw code during non-measurement periods / independent channels; recording the frequency of each unit using a counter array; calculating the total number of samples and the system period; calculating the calibration results of each unit, as well as differential and integral nonlinear indices; calculating the correction time value based on the indices; storing the calibration time in on-chip storage using the raw code as an index; constructing a calibration lookup table; during actual detection, using the raw code to address and look up the table, reading the calibration time and combining it with the coarse count to synthesize a high-precision result. This method can improve the measurement accuracy, stability, and environmental adaptability of a high-precision time-to-digital converter.