3results about How to "Small absolute error" patented technology

The invention relates to a road alignment safety evaluation method based on three-dimensional point cloud spatial modeling, and aims to overcome the defects of poor adaptability of a road surface point cloud segmentation algorithm and strong dependency of vertical line fitting on a road marking in a complex scene by utilizing characteristics of obtained high-density point cloud data. The road key information collaborative extraction method comprises the following steps: firstly, reducing the influence of environmental noise on road surface features through a normalized gradient filtering algorithm and a statistical outlier filtering algorithm; secondly, constructing a joint clustering model and an interactive region growing algorithm considering elevation, normal vector and pavement material reflection heterogeneity, and realizing precise segmentation of pavement point cloud in a complex scene; and then extracting a road surface edge line, obtaining road center line data, and generating a safety evaluation result list based on road center line parameters of a road plane, a longitudinal section and a cross section line shape. The research results provide technical support for application scenes such as road maintenance detection and automatic driving high-precision map construction.

This invention discloses a noise perception annoyance prediction method based on a convolutional recurrent neural network, comprising the following steps: Step 1) generating a preprocessed monophonic noise signal sample set; Step 2) generating a time-frequency image; Step 3) constructing a training sample set and a validation sample set; Step 4) training a deep neural network model using the training sample set to obtain a noise perception annoyance prediction model; Step 5) fitting a univariate linear relationship between the actual perceived annoyance and the original network prediction value using the validation sample set to obtain a linear correction function; Step 6) obtaining the noise perception annoyance prediction value; Step 7) correcting the noise perception annoyance prediction value using the linear correction function to obtain the final noise perception annoyance prediction structure. This invention maintains a low absolute error and an average relative error of approximately 1% to 2% even on high-perceived-annoyance noise samples, providing a stable and reliable objective indicator for noise assessment.

The invention discloses a rock reservoir average porosity estimation method, electronic equipment and a storage medium, and the method comprises the steps: obtaining the maximum effective porosity, the minimum effective porosity and the average porosity of a plurality of existing well target reservoirs around a research work area, calculating the maximum and minimum arithmetic average porosity of the maximum effective porosity and the minimum effective porosity of the target reservoir of each existing well; based on the average porosity of a plurality of existing well target reservoirs and the maximum and minimum arithmetic average porosity, fitting to obtain correlation coefficients in the corresponding function relation formula; and substituting the calculated maximum and minimum arithmetic average porosity of the maximum effective porosity and the minimum effective porosity of the target reservoir in the research work area into the corresponding function relation formula, and calculating the average porosity of the target reservoir in the research work area. According to the method, the porosity of the target reservoir in the research work area can be accurately estimated under the condition that only the maximum effective porosity and the minimum effective porosity in the rock reservoir sample in the research work area exist.