115results about How to "The predicted value is accurate" patented technology

The invention discloses a depth map intra prediction method based on a linear model. The gray values and coordinates of the previous line and left row of adjacent pixels of a current coded block are utilized to determine a linear model parameter; and according to the parameter and the pixel coordinate of the current coded block, the pixel gray value of the current coded block is predicted. According to the spatial character of a depth map, the depth map intra prediction method based on a linear model has the advantages of accurate prediction; meanwhile, because the previous line and left row of adjacent pixels of the current coded block are adopted to calculate the model parameter, a coding end does not need to code the model parameter; and a decoding end can directly determine the model parameter. The depth map intra prediction method can be applied to the coding standard of a three-dimensional video.

The invention belongs to the technical field of road traffic detection and discloses a GPS data-based prediction method for predicting the traffic jam-resulted delay time of a bus. The method comprises the steps of pre-processing the historical GPS data of a bus; conducting the statistic analysis on the pre-processed historical GPS data of the bus; classifying the traffic jam state of the bus; discriminating the state of the bus; predicting and correcting the traffic jam-resulted delay time of the bus. According to the technical scheme of the invention, the advantage that a large volume of bus GPS data are available is fully utilized. Meanwhile, in combination with the historical statistical information and real-time GPS data, traffic jam states are classified into six conditions through the decision tree classification method and different prediction calculation methods are adopted for different conditions. Moreover, prediction results are corrected according to the real-time data and the historical traffic jam information. Therefore, more accurate prediction values can be obtained in recurrent and accidental traffic congestion conditions.

A neural network system for predicting time series data. The system receives analyzed data obtained by multiresolution analysis of the time series data. The input processing layer of the system includes a series of neurons corresponding to the different levels of analysis, each neuron receiving the analyzed data for it own level. The output of each of these neurons is supplied as an additional input to the neuron for the next lower level of analysis. A predicted value is derived from the output of the neuron at the lowest level. The passing of results from one level to another improves prediction accuracy and simplifies the structure of any further processing layers.

An information acquiring apparatus acquires, when a travel object such as a traveler travels with transportation, passage time at which the travel object passes through each passage point at a departure / arrival facility, transportation specifying information indicating transportation, situation information indicating a situation and so forth. The travel process prediction apparatus stores the acquired information in an associated manner, and obtains a regression equation representing the relationship between items included in the transportation specifying information or situation information and passage time at a specific passage point, elapsed time while the travel object passes through two specific passage points or a result of comparison between the passage time and boarding completion time. The travel process prediction apparatus calculates a predicted value of future passage time, elapsed time or a result of comparison by substituting the content of the expected transportation specifying information or situation information for the regression equation.

The invention discloses a method and a device for intra-frame prediction. The method comprises the following contents: defining a straight line according to the direction of the intra-frame prediction and the position of a to-be-forecasted pixel; selecting the surrounding pixel of A which is an intersection of the line and the coded pixel as a reference pixel; conducting the intra-frame prediction on the base of an original value of the reference pixel; and forecasting that an adopted filter coefficient is related to the distance between the A and the refer pixel. Through the design of the reasonable intra-frame prediction method, the pixel value of the intra-frame prediction is closer to the original pixel; the problems that the forecasting is inaccurate and the entropy coding is redundant, as only the simple forecasting technology is applied for a multi-direction intra-frame prediction are solved; and the more accurate forecasting value can be obtained. Therefore, the residual error of the forecasting is reduced, and the coding efficiency is improved.

The invention discloses a real-time digital image compression prediction method, facilitating realization of a digital circuit and capable of performing time sequence optimization, based on prediction. Based on the prediction method, according to the image resolution, realization difficulty of the digital circuit and trade-off of compression efficiency, numerical values of a group of unit and prediction interval are selected. Higher is the image resolution, higher is the requirement for a time sequence of the digital circuit, at the moment, and the numerical values of the unit and prediction interval can be appropriately increased; according to the values of the selected unit and prediction interval and the position of the unit where a current pixel is located, corresponding reference pixels are selected for calculating the prediction value of the current pixel. By the adoption of the prediction method, the requirement for the time sequence of the digital circuit in a chip can be lowered, high practical value is achieved, and the method is expected to be applied to integrated circuits needing real-time video stream compression and decompression.

The invention discloses a multi-temporal dimension data fusion method for the large data of a power distribution network. The method comprises the steps of classifying data according to data sources; setting a statistical cycle and determining a sliding window value; calculating the smoothing factor value and the predictive value of the n cycle; calculating the deviation degree between an actual monitoring value and the predictive value of the n cycle; comparing the deviation degree with a preset deviation degree; when the deviation degree is larger than the preset deviation degree, calculating and obtaining a corrected sliding window value; when the deviation degree is smaller than the preset deviation degree, obtaining actual monitoring values obtained during several cycles from the n cycle and ranking the monitoring values according to the numbers of the cycles to obtain a window data set; calculating the weight coefficient of each data; calculating to obtain reported data; adding the reported data obtained through calculation to a similar reporting data set; and forming a new data set based on similar reporting data sets obtained through calculation. According to the technical scheme of the invention, the sliding window value can be dynamically adjusted. Meanwhile, data within a window can be fused according to weight factors in real time, so that the better data fusion effect is ensured. Moreover, a data basis is provided for upper-layer services.