38 results about "Cirrus cloud" patented technology

The invention discloses a cirrus cloud detection method, which comprises the following steps: inputting a remote sensing image, and acquiring a primary color channel grayscale image of the remote sensing image; acquiring a fractal dimension characteristic map of the primary color channel grayscale image; based on the fractal dimension characteristic map and The primary color channel grayscale image obtains the weight grayscale image; performs clustering calculation on the pixels of the primary color channel grayscale image, obtains the clustering results of the pixels, and determines the prediction value of each pixel in the primary color channel grayscale image based on the clustering results Assign labels; establish a graph cut model based on the primary color channel grayscale image, weight grayscale image, and pre-assigned labels; perform minimum cut calculations on the graph cut model, and output the detection results of remote sensing images. The invention also discloses a cirrus cloud detection device and a computer-readable storage medium at the same time.

The invention discloses an infrared imaging cirrus cloud detection method based on fractal dictionary learning, which relates to the technical field of infrared image processing, because there are a large number of false alarm sources with high radiation energy in the infrared imaging band, and the false alarm sources will interfere with the target detection. The specific steps of the present invention for solving the problem of false alarm source interference target detection are as follows: 1. Input an infrared image of cirrus cloud to be detected, and filter and denoise it to obtain a preprocessed cirrus cloud image; 2. Obtain a preprocessed cirrus cloud image The Hurst exponent value of image; 3. get random fractal image based on the Hurst exponent value obtained in step 2 and the rhombus-square subdivision method, and construct an over-complete dictionary D based on random fractal image; 4. utilize K-SVD algorithm to process The complete dictionary D performs sparse coding and dictionary update to obtain a sparse representation image; 5. Performs morphological filtering and image segmentation processing on the sparse representation image to obtain a detection result map. The detection method is easy to implement, has high detection accuracy and short calculation time of the algorithm.

The invention discloses a fully polarized laser radar system for detecting particle morphology, comprising four parts: a polarized light generating system, a receiving system, a polarization analysis system and a data acquisition processor; the invention also discloses a system for detecting particle morphology The detection method of the fully polarized lidar system is characterized in that the method is specifically based on the steps of detecting the atmospheric backscattering Miller matrix by using the fully polarized lidar system, and then detecting the shape of the atmospheric particles by the atmospheric backscattering matrix. The fully polarized laser radar system for detecting particle shape of the present invention uses multi-directional linearly polarized laser and circularly polarized laser to detect aerosol particles, which can realize fine discrimination and detection of the shape of aerosol particles in the atmosphere; Fine measurements of cloud particles are used to analyze the ice-water phase state of cloud particles and particle orientation in cirrus clouds.

The invention discloses a cirrus cloud recognition method based on satellite 2.0-micrometer channel data. The cirrus cloud recognition method comprises the steps of simulating satellite observation spectral line simulation values of a 2.0-micrometer channel and an oxygen A belt channel free of scattering factors and with cirrus cloud and aerosol under different observation geometrical conditions through a forward model according to satellite data; analyzing statistical characteristics of the satellite observation spectral line simulation values of the 2.0-micrometer channel and the oxygen A belt channel, wherein the statistical characteristics include spectrum mean values and spectrum variances; reading data of the satellite oxygen A belt channel and the 2.0-micrometer channel to generate satellite actually-measured spectral line values; performing calculation on the statistical characteristics of two satellite actually-measured spectral line values, and contrasting statistical laws of the two satellite actually-measured spectral line values to recognize whether cirrus cloud exists in a satellite observation field of view: judging whether influence of scattering shadow exists according to the spectrum mean value of the satellite actually-measured spectral line value of the oxygen A belt channel and determining the cirrus cloud or the aerosol through the spectrum variance of the satellite actually-measured spectral line value of the 2.0-micrometer channel if influence of scattering shadow exists. By means of the cirrus cloud recognition method based on satellite 2.0-micrometer channel data, the problem of scattering influence of the cirrus cloud during CO2 remote sensing retrieval of a short-wave near-infrared satellite can be solved quickly and effectively.

The invention provides a method for realizing 3D cloud layer simulation, which comprises the following steps of: A. making a 3D sky dynamic background varying with time; B. drawing a plurality of design sketches for a cloud sheet to form cloud form images; C. generating a single cloudlet by arranging a plurality of cloud sheets in a cubic space, carrying out the random sampling from the cloud form images by using random setting of UV coordinates according to different heights of cloud layers; D. setting the position attribution of each cloudlet after generating a plurality of the single cloudlets so as to generate a cloud layer; E. regulating the display of the 3D dynamic sky background in real time and setting real-time colors of the cloud layers according to current time; F. adding the dynamic effect for the cloud layers. By adopting the method, efficient and complicated cloud layer effects can be made by using several 2D design sketches of cirrus clouds or cumulus clouds and regulating the quantity of the cloud sheets and clouds, then more real and natural cloud layers can be simulated by regulation of a light source and addition of the dynamic effects.

The invention discloses a sun photometer measuring cirrus cloud optical characteristics; a host industrial control computer sends a measuring order, determines the initial position, and starts a sun visual locus tracking order; a program calculates horizontal and pitching motor steps needed to turn to the sun orientation, and sends the data to a lower computer embedded control system; the lower computer embedded control system controls a motor-driven system to rotate the horizontal and pitching motors, thus enabling a probe to turn to the current sun orientation; when the sun image appears ina CCD probe visual field, CCD image fine tracking is started; the sun lights pass the tracking imaging optical system, and irradiate on the target surface center of the CCD probe; the sun images in the visual field are gathered and stored, and sent to the host industrial control computer; the host industrial control computer analyzes the gathered images, thus determining whether the sun is finelytracked or not. The sun photometer employs the image tracking technology to replace the four-quadrant tracking, thus realizing all-weather tracking, and realizing large field of view tracking.