9 results about "Logarithmic scale" patented technology

The invention relates to the field of robot balance control, and discloses a control parameter optimization method of a second-order balance robot. According to the method, for a second-order balance system formed by connecting a vehicle body and a swing rod in series, a dynamic model and a linearization discrete state space model are established, and a linear quadratic regulator control framework is constructed; taking diagonal elements of the state weight matrix as to-be-optimized parameters, mapping the to-be-optimized parameters to a logarithmic scale search space after left-right symmetry constraint dimensionality reduction processing, and performing optimization by adopting a particle swarm optimization algorithm; and a composite cost function including a core error term, a control saturation term and a control stability term is constructed based on vehicle body inclination angle response, swing rod inclination angle response, control input and closed-loop poles, and an optimal state weight matrix and a corresponding control gain are output to serve as optimal control parameters of the second-order balance robot. The second-order balance robot control parameter setting method can improve pertinence and stability of second-order balance robot control parameter setting.

The utility model discloses a rapid navigation plotting ruler, which relates to the technical field of navigation plotting and comprises a ship position line plotting ruler and a plotting measuring ruler. Wherein a fan-shaped pattern with a vertex angle of 90 degrees is arranged on a panel of the ship position line plotting ruler; first open grooves are formed in two radiuses of the fan-shaped pattern, an open hole is formed in the circle center of the fan-shaped pattern, and angle scales are arranged at the arc of the fan-shaped pattern; a speed measuring ruler pattern and a drawing ruler pattern are arranged on a panel of the drawing measuring ruler; a second open groove is formed in the drawing ruler pattern; logarithmic scale scales are arranged on the edge of the plotting measuring scale. According to the utility model, the technical problem that the existing navigation plotting process is inconvenient to operate is solved.

The application discloses a kind of unmanned aerial vehicle micro target detection methods based on dynamic loss scheduling, belongs to computer vision target detection technical field, to solve the problem that the size of unmanned aerial vehicle aerial micro target is small, background is complex, shallow feature is easy to introduce false alarm, small batch mixed precision training is easy to shock and collapse.This method embeds P2 shallow high-resolution features as an implementation carrier in the target detection network, constructs a SIL logarithmic scale penalty term in the bounding box regression loss, and combines Inner-IoU, momentum smoothing focus and DLS three-stage dynamic loss scheduling;At the same time, through floating-point promotion, denominator lower bound truncation and exponential upper limit fuse, a numerical anti-collapse pipeline is formed to improve the positioning accuracy, training stability and engineering feasibility of micro targets, which can be used for unmanned aerial vehicle inspection, traffic monitoring, security patrol and low-altitude remote sensing target detection.

The application discloses a method and system for quickly determining a hydrocarbon generation threshold depth of argillaceous hydrocarbon source rock, and the method comprises the following steps: calculating total porosity of the argillaceous hydrocarbon source rock according to conventional logging data of an oil and gas exploration area; performing logarithmic scaling on the total porosity of the argillaceous hydrocarbon source rock to obtain a total porosity curve of the argillaceous hydrocarbon source rock; performing logarithmic scaling on the resistivity of the argillaceous hydrocarbon source rock to obtain a resistivity curve of the argillaceous hydrocarbon source rock; and determining the hydrocarbon generation threshold depth of the argillaceous hydrocarbon source rock according to the logarithmic scaling superposition result of the total porosity curve and the resistivity curve of the argillaceous hydrocarbon source rock. The method and system for determining the hydrocarbon generation threshold depth of the argillaceous hydrocarbon source rock can quickly determine the hydrocarbon generation threshold depth of the argillaceous hydrocarbon source rock in an exploration target area, so that the purpose of quickly determining the hydrocarbon generation threshold depth of the argillaceous hydrocarbon source rock at extremely low cost is achieved, and the efficiency of oil and gas exploration is improved.

This invention relates to the field of UAV image processing technology, specifically to a method and system for diagnosing secondary degradation of alpine grasslands based on multi-source data. Existing technologies suffer from the following drawbacks: low image processing accuracy, large data processing errors, and a lack of spatiotemporal comparability. To address these shortcomings, this invention first acquires images using the hypotenuse target method and calculates the modulation transfer function characteristic parameters to construct a multi-scale Gaussian space. An adaptive threshold algorithm is used to determine and fix thresholds, achieving binarization of the multi-scale images. Connected components meeting certain conditions are extracted from the binary images at each scale, and their directional robust perimeters are calculated, forming a perimeter sequence. Based on the perimeter sequence, a logarithmic scale-perimeter curve is established, and the second derivative is calculated to determine the collapse scale and standard scale, calculating the geometric features of the connected components. Combining degradation judgment thresholds and proportional constraints, crack-type early degradation patches satisfying the condition with the maximum negative second-order curvature are identified, achieving automatic identification of secondary grassland degradation.

According to the multi-frequency-point non-integer periodic disturbance signal generation and control method and system, reference frequencies are generated based on a multi-layer neural network, and a candidate frequency set is formed by the reference frequencies and integer multiples of the reference frequencies; the perceptron distributes the candidate frequency of which the state coefficient is 1 as the disturbance frequency; determining an amplitude coefficient in an amplitude range by adopting a single-layer neural network; grouping the disturbance frequencies according to the reference frequency; taking uniform distribution of the disturbance frequency on the logarithmic scale, minimum capacity required for injecting the signal at the disturbance frequency and uniform distribution of the disturbance frequency in each group as optimization objectives to obtain multiple groups of disturbance signals; determining dq-axis reference values of each group of disturbance signals injected into the power grid by adopting a reinforcement learning method by taking maximization of the direct-current voltage utilization rate and minimization of the capacity of disturbance injection equipment as targets; a discrete controller of disturbance signals is formed; the discrete controller is integrated in a control system of the disturbance injection device. Therefore, the capacity required by the injection disturbance is smaller, the DC voltage utilization rate is higher, and the disturbance signal is accurately controlled.

A semiconductor diode device according to the present invention comprises a guard ring layer positioned at a first surface side of a semiconductor substrate in a terminus region. The p-type impurity concentration in the guard ring layer is distributed so as to have a first curve for which the slope in a logarithmic scale gradually changes from a first surface to a first depth, a second curve for which the slope in the logarithmic scale gradually changes from the first depth to a second depth, which is the bottom of the guard ring layer, and a second inflection point at which the direction of the slope is reversed at the second depth such that the slope straddles the minimum value of the concentration. The terminus of the first curve and a start end of the second curve are connected at the first depth, and the p-type impurity concentration is distributed so as to have a first inflection point at the first depth at which the curves are bent such that the slope of the start end of the second curve becomes smaller than the slope of the terminus of the first curve.

The invention discloses a model training method and device, electronic equipment and a computer program product. The method comprises the following steps: constructing a target model for monocular depth estimation; obtaining training data including a color image and a corresponding real depth map; processing the color image based on the target model to obtain a predicted depth map; a preset mixed loss function is adopted to calculate the loss between the predicted depth map and the real depth map, parameters of the target model are adjusted according to the loss, and the mixed loss function is the weighted sum of a scale-invariant loss item, a gradient loss item and a smooth loss item; the scale invariant loss item is obtained by predicting the difference between the depth map and the real depth map on the logarithmic scale, the gradient loss item is obtained by predicting the difference between the depth map and the real depth map on the gradient space, and the smooth loss item is obtained by performing weighted punishment on the gradient of the depth map through the gradient of the color image. According to the scheme, the quality of the depth map output by the obtained depth estimation model can be improved.

The application discloses a kind of based on double-flow self-adapting physical information fusion forging press energy consumption prediction system and method, comprising: based on physical principle to establish energy consumption mechanism model, and according to real-time state parameter calculation each component theoretical energy consumption value as mechanism data, mechanism data and real-time state parameter are selected for feature, respectively construct mechanism dataset and acquisition dataset, mechanism flow uses time series encoder based on mechanism dataset to carry out prediction, and introduce adaptive correction mechanism to compensate non-modeling energy consumption, while using logarithmic scale soft constraint to keep reasonable deviation of predicted value and theoretical value;Data flow uses time series encoder based on acquisition dataset to carry out prediction driven by pure data, and is trained by minimizing its prediction error, based on Monte Carlo Dropout quantification double-flow prediction uncertainty, generate dynamic fusion weight, and under the constraint of anti-degeneration regularization, the weighted fusion of two-flow predicted values is carried out, and the final comprehensive energy consumption predicted value is output.