CN112419407A - Cloud cluster displacement vector calculation method and device based on cloud cluster edge identification - Google Patents

Cloud cluster displacement vector calculation method and device based on cloud cluster edge identification Download PDF

Info

Publication number
CN112419407A
CN112419407A CN202011218307.0A CN202011218307A CN112419407A CN 112419407 A CN112419407 A CN 112419407A CN 202011218307 A CN202011218307 A CN 202011218307A CN 112419407 A CN112419407 A CN 112419407A
Authority
CN
China
Prior art keywords
edge
point
cloud
cloud cluster
adjacent pixel
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202011218307.0A
Other languages
Chinese (zh)
Other versions
CN112419407B (en
Inventor
梅生伟
张雪敏
周子杰
甄钊
杨滢璇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tsinghua University
State Grid Corp of China SGCC
China Electric Power Research Institute Co Ltd CEPRI
Electric Power Research Institute of State Grid Liaoning Electric Power Co Ltd
Original Assignee
Tsinghua University
State Grid Corp of China SGCC
China Electric Power Research Institute Co Ltd CEPRI
Electric Power Research Institute of State Grid Liaoning Electric Power Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tsinghua University, State Grid Corp of China SGCC, China Electric Power Research Institute Co Ltd CEPRI, Electric Power Research Institute of State Grid Liaoning Electric Power Co Ltd filed Critical Tsinghua University
Priority to CN202011218307.0A priority Critical patent/CN112419407B/en
Publication of CN112419407A publication Critical patent/CN112419407A/en
Application granted granted Critical
Publication of CN112419407B publication Critical patent/CN112419407B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/20—Image enhancement or restoration using local operators
    • G06T5/30—Erosion or dilatation, e.g. thinning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/13—Edge detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/66—Analysis of geometric attributes of image moments or centre of gravity
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10028—Range image; Depth image; 3D point clouds
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Geometry (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

The embodiment of the invention provides a cloud cluster displacement vector calculation method and a cloud cluster displacement vector calculation device based on cloud cluster edge identification, wherein the method comprises the following steps: dividing a cloud cluster area from the current cloud picture; extracting the outer edge of each cloud cluster region; determining the central point of each block region distributed along the outer edge on the outer edge of each cloud cluster region, and determining the position of each block region according to the central point of each block region; and calculating the displacement vector of each block area based on a block matching algorithm according to the position of each block area. The embodiment of the invention enables the displacement vector information of different positions of each cloud cluster to be expressed and improves the calculation efficiency and precision.

Description

Cloud cluster displacement vector calculation method and device based on cloud cluster edge identification
Technical Field
The invention relates to the technical field of displacement vector calculation, in particular to a cloud cluster displacement vector calculation method and device based on cloud cluster edge identification.
Background
The photovoltaic power prediction mainly comprises four classification modes, namely a prediction time scale, a prediction space scale, a prediction process and a prediction method. The blocking of clouds is a major factor affecting photovoltaic output. When a cloud cluster exists above the photovoltaic power station and is shielded, partial illumination cannot pass through a cloud layer, and then solar irradiance received by a photovoltaic panel of the power station is reduced, so that the photovoltaic power of the power station suddenly drops. The movement and the generative deformation of the cloud cluster are one of the root causes of uncertain fluctuation of photovoltaic output. The cloud cluster displacement vector is calculated and predicted based on the cloud picture, and the method is an important link for photovoltaic power prediction.
The current main cloud cluster displacement vector calculation method comprises two steps of cloud cluster identification and cloud displacement vector calculation. The main purpose of cloud cluster identification is to distinguish the pixels representing the cloud cluster from the background, thereby eliminating the interference of the brightness difference of the pixels on the calculation of the displacement vector, which is essentially a process of converting a gray-scale image or a color image into a binary image. The cloud displacement vector calculation is a core link of the algorithm, and because cloud clusters have deformation in the motion process, the traditional motion vector calculation method, such as the calculation of the gravity center moving distance, is not suitable any more.
At present, cloud motion vector calculation methods based on cloud image data mainly comprise a block matching method, a Lucas-Kanade optical flow method and a phase correlation algorithm. The block matching method is a common method for motion estimation, and is widely applied to motion analysis of two-dimensional image data. The basic principle is that based on a certain size of sub-area in the current frame, according to a certain measurement method, the best matching sub-area is searched in a certain range of the previous frame and the next frame of the current frame. Among them, Cross-correlation Coefficient (CCC), Sum of Absolute Differences (SAVD), and Sum of square Differences (Sum of Squared Differences (SSD) are three commonly used measurement methods. The Lucas-Kanade optical flow method is a differential method for optical flow estimation, calculates the movement of each pixel point position of two frames of images in the time from t to t + delta t, and is widely applied to image speed calculation. The phase correlation algorithm is an image registration method that exploits the properties of the fourier transform. After the characteristic matrix of the image is subjected to Fourier transform, the correlation operation on the time domain can be converted into the operation on the frequency domain. When the image is translated, the corresponding amplitude spectrum is unchanged, and the phase correlation algorithm is provided according to the principle.
The three cloud displacement vector calculation methods have three defects. First, the calculation efficiency is low, the displacement vector calculation is performed on the cloud map universe in the three current calculation methods, the calculated data volume is relatively large, especially for satellite cloud maps with numerous pixel points, the time for analyzing one cloud map is extremely long, and the requirement for the real-time property of ultra-short-term prediction cannot be met. Secondly, the calculation precision is low, and because the cloud cluster has the phenomenon of deformation during the motion process, and the three algorithms cannot accurately position the position of the cloud cluster when calculating the displacement vector, the obtained displacement vector result often has larger deviation from the actual result. Thirdly, the obtained information quantity is small, the Lucas-Kanade optical flow method and the phase correlation algorithm obtain the displacement vector result of the whole cloud picture, the block matching method obtains the displacement vector value of each closely-arranged block region of the cloud picture, the cloud groups in the cloud picture are randomly distributed, the speeds of the cloud groups and different regions of the same cloud group are different, and the information obtained by the current algorithm cannot reflect the phenomenon.
Disclosure of Invention
The embodiment of the invention provides a cloud cluster displacement vector calculation method and device based on cloud cluster edge identification, which are used for solving the defects of low efficiency and low precision of cloud displacement vector calculation and small amount of obtained information in the prior art, realizing the acquisition of displacement vectors of different positions of a cloud cluster, and improving the calculation efficiency and precision.
The embodiment of the invention provides a cloud cluster displacement vector calculation method based on cloud cluster edge identification, which comprises the following steps:
dividing a cloud cluster area from the current cloud picture;
extracting the outer edge of each cloud cluster region;
determining the central point of each block region distributed along the outer edge on the outer edge of each cloud cluster region, and determining the position of each block region according to the central point of each block region;
and calculating the displacement vector of each block area based on a block matching algorithm according to the position of each block area.
According to one embodiment of the invention, the cloud cluster displacement vector calculation method based on cloud cluster edge identification is used for segmenting a cloud cluster region from a current cloud picture, and comprises the following steps:
segmenting the current cloud picture based on a threshold segmentation method to obtain an initial cloud cluster in the current cloud picture;
performing expansion operation on the initial cloud cluster based on the current expansion coefficient to generate a new cloud cluster;
counting the number of new clouds with the number of pixels larger than a preset pixel number threshold value, and if the number of the new clouds is smaller than the preset cloud number threshold value, increasing the current expansion coefficient;
performing a dilation operation on the initial cloud cluster again by using the increased current dilation coefficient until the number of the generated new cloud clusters is greater than or equal to the preset cloud cluster number threshold;
and taking the new cloud cluster generated in the last iteration as the cloud cluster area segmented from the current cloud picture.
According to one embodiment of the invention, the cloud displacement vector calculation method based on cloud edge identification comprises the following steps of:
selecting any highest point from the outer edge of each cloud cluster area as an initial point for searching the outer edge of each cloud cluster area;
for the edge point of any cloud cluster area searched last time, searching a next edge point of the cloud cluster area according to the edge type of the edge point and the pixel point adjacent to the edge point, and determining the edge type of the next edge point until the next edge point is overlapped with the initial point of the outer edge of the cloud cluster area and the iteration number is more than 3;
wherein the edge types include an upper edge, a lower edge, a left edge, and a right edge; the upper side adjacent pixel point of the edge point with the edge type of the upper edge is the background of the current cloud picture, the lower side adjacent pixel point of the edge point with the edge type of the lower edge is the background, the left side adjacent pixel point of the edge point with the edge type of the left edge is the background, and the right side adjacent pixel point of the edge point with the edge type of the right edge is the background;
and taking all the found edge points of the cloud cluster area as pixel points of the outer edge of the cloud cluster area.
According to an embodiment of the present invention, a cloud cluster displacement vector calculation method based on cloud cluster edge recognition, where the searching for a next edge point of a cloud cluster region according to an edge type of an edge point and a pixel point adjacent to the edge point and determining an edge type of the next edge point includes:
if the edge type of the edge point is an upper edge and the adjacent pixel point at the upper right corner of the edge point is a pixel point in the cloud cluster region, taking the adjacent pixel point at the upper right corner of the edge point as the next edge point, and taking the edge type of the next edge point as a left edge point;
if the edge type of the edge point is an upper edge, an adjacent pixel point at the upper right corner of the edge point is the background, and an adjacent pixel point at the right side is a pixel point in the cloud cluster region, taking the adjacent pixel point at the right side of the edge point as the next edge point, and taking the edge type of the next edge point as the upper edge point;
if the edge type of the edge point is an upper edge, and the adjacent pixel point at the upper right corner and the adjacent pixel point at the right side of the edge point are the backgrounds, taking the edge point as the next edge point, and taking the edge type of the next edge point as a right edge point;
if the edge type of the edge point is a lower edge and the adjacent pixel point at the lower left corner of the edge point is a pixel point in the cloud cluster area, taking the adjacent pixel point at the lower left corner of the edge point as the lower edge point, and taking the edge type of the lower edge point as a right edge point;
if the edge type of the edge point is a lower edge, the adjacent pixel point at the lower left corner of the edge point is the background, and the adjacent pixel point at the left side is the pixel point in the cloud cluster area, taking the adjacent pixel point at the left side of the edge point as the next edge point, and taking the edge type of the next edge point as the lower edge point;
if the edge type of the edge point is a lower edge, and the adjacent pixel point at the lower left corner and the adjacent pixel point at the left side of the edge point are the backgrounds, taking the edge point as the lower edge point, and taking the edge type of the lower edge point as a left edge point;
if the edge type of the edge point is a left edge and the adjacent pixel point at the upper left corner of the edge point is a pixel point in the cloud cluster region, taking the adjacent pixel point at the upper left corner of the edge point as the next edge point, and taking the edge type of the next edge point as the lower edge point;
if the edge type of the edge point is a left edge, an adjacent pixel point at the upper left corner of the edge point is the background, and an adjacent pixel point at the upper side is a pixel point in the cloud cluster region, taking the adjacent pixel point at the left side of the edge point as the next edge point, and taking the edge type of the next edge point as the left edge point;
if the edge type of the edge point is a left edge, and the adjacent pixel point at the upper left corner and the adjacent pixel point at the upper side of the edge point are the backgrounds, taking the edge point as the next edge point, and taking the edge type of the next edge point as the upper edge point;
if the edge type of the edge point is the right edge and the adjacent pixel point at the lower right corner of the edge point is the pixel point in the cloud cluster area, taking the adjacent pixel point at the upper right corner of the edge point as the next edge point, and taking the edge type of the next edge point as the upper edge point;
if the edge type of the edge point is the right edge, the adjacent pixel point at the lower right corner of the edge point is the background, and the adjacent pixel point at the lower side is the pixel point in the cloud cluster area, taking the adjacent pixel point at the lower side of the edge point as the next edge point, and taking the edge type of the next edge point as the right edge point;
and if the edge type of the edge point is the right edge, and the adjacent pixel point at the lower right corner and the adjacent pixel point at the lower side of the edge point are the backgrounds, taking the edge point as the next edge point, wherein the edge type of the next edge point is the lower edge point.
According to an embodiment of the invention, the cloud displacement vector calculation method based on cloud edge identification, wherein the step of determining the central point of each block region distributed along the outer edge on the outer edge of each cloud region comprises:
sequentially sequencing the pixel points on the outer edge of each cloud cluster area along a clockwise sequence;
for the central point determined from the sorting result last time, retrieving a pixel point set from the sorting result, wherein the difference between the abscissa and the abscissa of the central point determined last time is equal to the side length of the block area, or the difference between the ordinate and the ordinate of the central point determined last time is equal to the side length of the block area;
and selecting the pixel point which is positioned behind the last determined central point in the sequencing result from the pixel point set, and taking the pixel point which is closest to the last determined central point in the selected pixel points as the central point of the block region until the closest pixel point is the last pixel point in the sequencing result.
According to the cloud displacement vector calculation method based on cloud edge identification, the center point of each block region distributed along the outer edge is determined on the outer edge of each cloud region, and then the method further comprises the following steps:
sequencing the central points according to the sequence determined by the central points;
sequentially selecting a central point from the sequencing results of the central points;
calculating the overlapping degree between the block region where the central point is located and the block region where the selected central point is located for any central point behind the selected central point;
and if the overlapping degree is greater than a preset overlapping degree threshold value, deleting the central point from the sequencing result of the central points.
According to the cloud displacement vector calculation method based on cloud edge identification, the center point of each block region distributed along the outer edge is determined on the outer edge of each cloud region, and then the method further comprises the following steps:
modifying the position of the central point of each block area according to the current expansion coefficient of the last expansion operation on the initial cloud cluster and the edge type of the central point of each block area;
wherein the edge types include an upper edge, a lower edge, a left edge, and a right edge; the upper side adjacent pixel point with the edge type being the central point of the upper edge is the background of the current cloud picture, the lower side adjacent pixel point with the edge type being the central point of the lower edge is the background, the left side adjacent pixel point with the edge type being the central point of the left edge is the background, and the right side adjacent pixel point with the edge type being the central point of the right edge is the background.
The embodiment of the present invention further provides a cloud cluster displacement vector calculation apparatus based on cloud cluster edge identification, including:
the segmentation module is used for segmenting cloud cluster areas from the current cloud picture;
the extraction module is used for extracting the outer edge of each cloud cluster area;
the determining module is used for determining the central point of each block region distributed along the outer edge on the outer edge of each cloud cluster region and determining the position of each block region according to the central point of each block region;
and the calculation module is used for calculating the displacement vector of each block area based on a block matching algorithm according to the position of each block area.
The embodiment of the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of any one of the cloud edge identification-based cloud displacement vector calculation methods described above when executing the program.
Embodiments of the present invention further provide a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of the cloud displacement vector calculation method based on cloud edge identification as described in any one of the above.
According to the cloud cluster displacement vector calculation method and device based on cloud cluster edge recognition, the outer edge of each cloud cluster region in a cloud cluster is extracted, the central point of each block region is determined on the outer edge of each cloud cluster region, and the displacement vector of each block region is calculated based on a block matching algorithm; on the other hand, the displacement vectors at the positions on the outer edge of the cloud cluster are obtained through a block matching algorithm, so that the displacement vector information of different positions of the cloud cluster can be fully expressed, and the accuracy of cloud cluster displacement vector calculation is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
Fig. 1 is a schematic flowchart of a cloud displacement vector calculation method based on cloud edge identification according to an embodiment of the present invention;
fig. 2 is a schematic view of a complete flow of a cloud displacement vector calculation method based on cloud edge identification according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a cloud displacement vector calculation apparatus based on cloud edge identification according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
A cloud displacement vector calculation method based on cloud edge identification according to an embodiment of the present invention is described below with reference to fig. 1, where the method includes: step 101, dividing a cloud cluster area from a current cloud picture;
wherein, the current cloud picture is the cloud picture which needs to be subjected to displacement vector calculation. The cloud picture comprises a cloud cluster area and a background area. And segmenting the cloud cluster region from the current cloud picture by using a segmentation algorithm. The present embodiment is not limited to a specific segmentation method.
Step 102, extracting the outer edge of each cloud cluster area;
the segmentation result of the current cloud picture can be regarded as a binary cloud picture. Wherein, the pixel values of the background area in the binary cloud picture are the same and are 0; the pixel values of the cloud cluster regions are all the same and are 1. Therefore, the binary cloud picture cannot reflect the motion information of the cloud cluster, and only the edge motion information of the cloud cluster is valuable. Therefore, the present embodiment extracts the outer edge of each cloud region.
103, determining the central point of each block area distributed along the outer edge on the outer edge of each cloud cluster area, and determining the position of each block area according to the central point of each block area;
after the outer edge of each cloud cluster area is obtained, an edge point is selected from the outer edge as the center point of the block area. The present embodiment is not limited to a specific selection method. The selection of the center point of the patch area causes the patch area to be evenly distributed on the outer edge of the cloud cluster area.
And 104, calculating the displacement vector of each block area based on a block matching algorithm according to the position of each block area.
The block matching algorithm is based on a partition area with a certain size in the current cloud picture, and an optimal matching partition area is searched in a certain range of frames before and after the current cloud picture according to a certain measurement method. The present embodiment does not limit the measurement method. And the displacement vector of each block area on the edge of each cloud image area is the displacement vector of the edge of each cloud image area.
In the embodiment, the outer edge of each cloud image region in the cloud image is extracted, the central point of each block region is determined on the outer edge of each cloud cluster region, and the displacement vector of each block region is calculated based on the block matching algorithm, so that on one hand, only the edge motion information of the cloud cluster is valuable for calculating the cloud image displacement vector, and therefore, the displacement vector is calculated only for each block region on the cloud cluster edge, and the calculation efficiency is improved; on the other hand, the displacement vectors at the positions on the outer edge of the cloud cluster are obtained through a block matching algorithm, so that the displacement vector information of different positions of the cloud cluster can be fully expressed, and the accuracy of cloud cluster displacement vector calculation is improved.
On the basis of the foregoing embodiment, the segmenting the cloud cluster region from the current cloud map in this embodiment includes: segmenting the current cloud picture based on a threshold segmentation method to obtain an initial cloud cluster in the current cloud picture;
the threshold segmentation method is characterized in that a corresponding threshold is set according to the difference of certain parameter values of a cloud cluster area and a background area in a current cloud picture so as to distinguish the cloud cluster area from the background area. The common threshold segmentation method is a maximum inter-class variance method, the current cloud picture is divided into two parts, namely a cloud cluster area and a background area, according to the gray level difference of each pixel in the current cloud picture, when the inter-class variance of the cloud cluster area and the background area reaches the maximum value, the difference between the two parts is most obvious, the segmentation is most accurate at the moment, and the threshold with the maximum inter-class variance is the set threshold. The inter-class variance g is calculated as follows:
g＝w1(μ1-μ)2+w2(μ2-μ)2；
in the formula, w1And mu1Is the pixel point proportion and the average gray value, w, of the cloud cluster region2And mu2The ratio and the average gray value of the pixel points in the background area are shown, and mu is the average gray value of all the pixel points in the current cloud picture. In this embodiment, the cloud cluster segmented from the current cloud image by using the threshold segmentation method is used as the initial cloud cluster.
Performing expansion operation on the initial cloud cluster based on the current expansion coefficient to generate a new cloud cluster; counting the number of new clouds with the number of pixels larger than a preset pixel number threshold value, and if the number of the new clouds is smaller than the preset cloud number threshold value, increasing the current expansion coefficient; performing a dilation operation on the initial cloud cluster again by using the increased current dilation coefficient until the number of the generated new cloud clusters is greater than or equal to the preset cloud cluster number threshold; and taking the new cloud cluster generated in the last iteration as the cloud cluster area segmented from the current cloud picture.
In addition to large clump clouds, current clouds tend to be populated with a large number of fragmented small clouds. If the displacement vector calculation is performed on all the clouds separately, the calculation efficiency is greatly reduced. In fact, closely located clouds tend to have similar velocities and trends in motion. Therefore, the cloud clusters with similar geographic positions can be regarded as a whole, and the moving speed of different small piece clouds at the position can be approximately represented by calculating the moving speed of each position of the whole cloud cluster, so that the computing efficiency is effectively improved, and the computing precision is ensured.
The embodiment divides the cloud cluster area according to the physical interval. And when the minimum distance between different clouds is smaller than a set distance threshold value, regarding the clouds as the same cloud cluster area. The method for judging whether the cloud cluster distance is smaller than the threshold value is realized through expansion operation. In the imaging, the cloud cluster can be represented as a two-dimensional matrix, the expansion operation can realize the uniform expansion of the two-dimensional matrix towards the periphery, and the expansion size is determined by the expansion coefficient width. When the two expanded cloud cluster matrixes are connected, the two expanded cloud cluster matrixes are divided into the same cloud cluster area.
The specific algorithm is as follows:
step 1, inputting: processing by a maximum inter-class variance method to obtain a binary cloud picture;
step 2, initialization: setting the initial expansion coefficient width to be 1;
step 3, expanding width pixels of all clouds in the current cloud picture;
step 4, counting the number of the pixels in the cloud picture to be larger than a preset pixel number threshold value S0The number of all cloud cluster areas is marked as m;
step 5, when m is<Presetting a cloud cluster number threshold m0And (5) turning to the step (3), wherein the width is equal to width + 1; when m is more than or equal to m0Go to step 6;
and 6, outputting the expansion coefficient width and the cloud cluster area division result.
On the basis of the foregoing embodiment, the extracting the outer edge of each cloud cluster region in this embodiment includes: selecting any highest point from the outer edge of each cloud cluster area as an initial point for searching the outer edge of each cloud cluster area; for the edge point of any cloud cluster area searched last time, searching a next edge point of the cloud cluster area according to the edge type of the edge point and the pixel point adjacent to the edge point, and determining the edge type of the next edge point until the next edge point is overlapped with the initial point of the outer edge of the cloud cluster area and the iteration number is more than 3; wherein the edge types include an upper edge, a lower edge, a left edge, and a right edge; the upper side adjacent pixel point of the edge point with the edge type of the upper edge is the background of the current cloud picture, the lower side adjacent pixel point of the edge point with the edge type of the lower edge is the background, the left side adjacent pixel point of the edge point with the edge type of the left edge is the background, and the right side adjacent pixel point of the edge point with the edge type of the right edge is the background; and taking all the found edge points of the cloud cluster area as pixel points of the outer edge of the cloud cluster area.
In particular, for the extraction of the outer edge of each cloud region, it is essentially a question of finding a closed curve. Due to the existence of the cavity in the cloud cluster region, the existing method for searching the closed curve, such as a depth-first algorithm and a 4 or 8 connected domain method, is not suitable. Therefore, the present embodiment finds the outer edge of the cloud cluster region based on the edge type determination policy, and the edge type determination policy finds the next edge point for the cloud cluster pixel or the background pixel according to the edge type of the edge point and the pixel point adjacent to the edge point.
The specific algorithm is as follows:
step 1, initialization: selecting any highest point (x) of the outer edge0,y0) As an initial point of edge search, defining the edge type of the point as an upper edge point;
step 2, based on point (x)i,yi) Find the next edge point (x) according to the edge type of the pointi+1,yi+1)；
Step 3, judging the strategy judgment point (x) based on the edge typei+1,yi+1) The edge type of (2);
step 4, when i is less than 3, turning to step 2; when i is more than or equal to 3, turning to the step 5;
step 5, judge the point (x)i+1,yi+1) And point (x)0,y0) Whether they coincide. When x isi+1＝x0，yi+1＝y0Go to step 6; otherwise, turning to the step 2, wherein i is i + 1;
step 6, outputting a pixel point matrix of the outer edge of the cloud cluster region as E ═ x0,x1,...,xn)T,(y0,y1,...,yn)T]。
On the basis of the foregoing embodiment, in this embodiment, finding a next edge point of the cloud cluster region according to the edge type of the edge point and a pixel point adjacent to the edge point, and determining the edge type of the next edge point includes: if the edge type of the edge point is an upper edge and the adjacent pixel point at the upper right corner of the edge point is a pixel point in the cloud cluster region, taking the adjacent pixel point at the upper right corner of the edge point as the next edge point, and taking the edge type of the next edge point as a left edge point;
the edge type determination policy is shown in tables 1 to 4, and describes all possible distribution situations of the pixel values around the edge points of different types of the cloud cluster and the determination policy to be adopted. Taking the above case of the edge point as an example, let the coordinates of the edge point be (x, y), and the coordinates of the upper-right adjacent pixel point of the edge point be (x +1, y-1). If the pixel value at (x +1, y-1) is 1, no matter whether the pixel value of the right-side neighboring pixel (x +1, y) is 0 or 1, the position of the next edge point is the upper right corner (x +1, y-1), and the edge type of the next edge point is the left edge point, as shown in the first and second rows in table 1. Wherein, 1 in table 1 represents a pixel point in the cloud cluster region, and 0 represents a background pixel point.
If the edge type of the edge point is an upper edge, and an adjacent pixel point at the upper right corner of the edge point is the background and an adjacent pixel point at the right side is a pixel point in the cloud cluster region, taking the adjacent pixel point at the right side of the edge point as the next edge point, and taking the edge type of the next edge point as the upper edge point, as shown in the third row in table 1;
if the edge type of the edge point is an upper edge, and the adjacent pixel point at the upper right corner and the adjacent pixel point at the right side of the edge point are the backgrounds, taking the edge point as the next edge point, and taking the edge type of the next edge point as a right edge point, as shown in the fourth line in table 1;
if the edge type of the edge point is a lower edge and the adjacent pixel point at the lower left corner of the edge point is a pixel point in the cloud cluster region, taking the adjacent pixel point at the lower left corner of the edge point as the lower edge point, and taking the edge type of the lower edge point as a right edge point, as shown in the first line and the third line in table 2;
if the edge type of the edge point is a lower edge, and the adjacent pixel point at the lower left corner of the edge point is the background and the adjacent pixel point at the left side is a pixel point in the cloud cluster region, taking the adjacent pixel point at the left side of the edge point as the lower edge point, and the edge type of the lower edge point is the lower edge point, as shown in the second row in table 2;
if the edge type of the edge point is a lower edge, and the adjacent pixel point at the lower left corner and the adjacent pixel point at the left side of the edge point are both the backgrounds, the edge point is taken as the lower edge point, and the edge type of the lower edge point is a left edge point, as shown in the fourth line in table 2;
if the edge type of the edge point is a left edge and the adjacent pixel point at the upper left corner of the edge point is a pixel point in the cloud cluster region, taking the adjacent pixel point at the upper left corner of the edge point as the next edge point, and taking the edge type of the next edge point as the lower edge point, as shown in the first row and the second row in table 3;
if the edge type of the edge point is a left edge, and an adjacent pixel point at the upper left corner of the edge point is the background and an adjacent pixel point at the upper side is a pixel point in the cloud cluster region, taking the adjacent pixel point at the left side of the edge point as the next edge point, and taking the edge type of the next edge point as the left edge point, as shown in the third row in table 3;
if the edge type of the edge point is the left edge, and the adjacent pixel point at the upper left corner and the adjacent pixel point at the upper side of the edge point are the backgrounds, the edge point is taken as the next edge point, and the edge type of the next edge point is the upper edge point, as shown in the fourth line in table 3;
if the edge type of the edge point is the right edge and the adjacent pixel point at the lower right corner of the edge point is the pixel point in the cloud cluster region, taking the adjacent pixel point at the upper right corner of the edge point as the next edge point, wherein the edge type of the next edge point is the upper edge point, as shown in the first line and the third line in table 4;
if the edge type of the edge point is the right edge, the adjacent pixel point at the lower right corner of the edge point is the background, and the adjacent pixel point at the lower side is the pixel point in the cloud cluster region, taking the adjacent pixel point at the lower side of the edge point as the next edge point, and the edge type of the next edge point is the right edge point, as shown in the second row in table 4;
if the edge type of the edge point is the right edge, and the adjacent pixel point at the lower right corner and the adjacent pixel point at the lower side of the edge point are both the backgrounds, the edge point is taken as the next edge point, and the edge type of the next edge point is the lower edge point, as shown in the fourth line in table 4.
In the cloud cluster edge type judgment strategy in this embodiment, different processing strategies are adopted for different types of edge points, so that the accurate search of the cloud cluster outer edge is realized.
TABLE 1 lower edge point finding of the upper edge
Figure BDA0002761173430000141
TABLE 2 Next edge Point finding of the lower edge
Figure BDA0002761173430000151
TABLE 3 Next edge Point finding of left edge
Figure BDA0002761173430000152
TABLE 4 Next edge Point finding of Right edge
Figure BDA0002761173430000153
On the basis of the foregoing embodiment, in this embodiment, the determining, on the outer edge of each cloud cluster region, the central points of the respective regions distributed along the outer edge includes: sequentially sequencing the pixel points on the outer edge of each cloud cluster area along a clockwise sequence; for the central point determined from the sorting result last time, retrieving a pixel point set from the sorting result, wherein the difference between the abscissa and the abscissa of the central point determined last time is equal to the side length of the block area, or the difference between the ordinate and the ordinate of the central point determined last time is equal to the side length of the block area;
specifically, after the coordinates of all pixel points on the outer edge of the cloud cluster region are obtained, the coordinates of the pixel points are sequentially arranged along a clockwise sequence, namely, the coordinates are stored in a two-dimensional matrix form. To distribute the block regions evenly over the outer edge of the cloud cluster region, the block regions should be exactly bordered, i.e. the difference between the abscissa or ordinate of the adjacent block regions is equal to the side length l of the block region. Therefore, as long as the initial coordinate point of the outer edge is used as a starting point, and pixel points with the difference between the abscissa and the ordinate equal to the side length of the block area are searched in the two-dimensional matrix in sequence, the coordinates of the center points of all the block areas can be determined in sequence.
And selecting the pixel point which is positioned behind the last determined central point in the sequencing result from the pixel point set, and taking the pixel point which is closest to the last determined central point in the selected pixel points as the central point of the block region until the closest pixel point is the last pixel point in the sequencing result.
The central point positioning algorithm of the block area comprises the following steps:
step 1, initialization: determining a starting point (x) in an outer edge coordinate matrix of a cloud cluster region0,y0) Is the central initial point of the block area;
step 2, searching and point (x) in the point coordinate matrix in sequencei,yi) Is equal to the set of point coordinates (x) of the side length l of the block areai1,yi1),(xi2,yi2),…,(xin,yin)；
Step 3, selecting the position number in the point coordinate set to be greater than (x)i,yi) The point coordinate of the minimum position number of (2) is recorded as the next block area center (x)i+1,yi+1)；
Step 4, when i +1<Number of matrix columns imaxIf so, i is i +1, and the step 2 is switched to; when i +1 is equal to the number of matrix columns imaxIf yes, turning to step 5;
step 5, obtaining a central point matrix F [ [ (x) of the block area0,x1,...,xq)T,(y0,y1,...,yq)T]。
On the basis of the foregoing embodiment, in this embodiment, the determining, on the outer edge of each cloud cluster region, the center points of the respective regions distributed along the outer edge, then further includes: sequencing the central points according to the sequence determined by the central points; sequentially selecting a central point from the sequencing results of the central points; calculating the overlapping degree between the block region where the central point is located and the block region where the selected central point is located for any central point behind the selected central point; and if the overlapping degree is greater than a preset overlapping degree threshold value, deleting the central point from the sequencing result of the central points.
Specifically, due to the irregular shape of the cloud cluster region, the actual distance between the central points of partial block regions is short, and the overlapping degree of the block regions is high. Therefore, the block area with higher overlapping degree needs to be processed, and when the overlapping rate of the block area and the block area is larger than a preset overlapping degree threshold value wmaxAnd deleting one of the areas. The steps of the overlap optimization are as follows:
step 1, determine (x) in F0,y0) An initial point optimized for overlap;
step 2, regarding the coordinate point (x)j,yj) Sequentially finding out points (x)j+1,yj+1)～(xm,ym) The overlapping degree w of the block areas and the block areas. When w is larger than the preset overlap threshold value wmaxThe coordinates of the point are deleted from the matrix, and the calculation formula of the overlap w is as follows:
Figure BDA0002761173430000171
step 3, when j is less than m-1, j is equal to j +1, and the step 2 is switched to; when j is m-1, go to step 4;
and 4, obtaining a block area central point matrix H after the overlapping degree is optimized.
On the basis of the foregoing embodiment, in this embodiment, the determining, on the outer edge of each cloud cluster region, the center points of the respective regions distributed along the outer edge, then further includes: modifying the position of the central point of each block area according to the current expansion coefficient of the last expansion operation on the initial cloud cluster and the edge type of the central point of each block area; wherein the edge types include an upper edge, a lower edge, a left edge, and a right edge; the upper side adjacent pixel point with the edge type being the central point of the upper edge is the background of the current cloud picture, the lower side adjacent pixel point with the edge type being the central point of the lower edge is the background, the left side adjacent pixel point with the edge type being the central point of the left edge is the background, and the right side adjacent pixel point with the edge type being the central point of the right edge is the background.
Specifically, since the current cloud image is subjected to the over-expansion operation using the expansion coefficient width, the position of the center of the block region needs to be corrected. For the
Figure BDA0002761173430000183
The following position corrections are performed:
Figure BDA0002761173430000181
Figure BDA0002761173430000182
obtaining a block region center point matrix H '[ (x'0,x'1,...,x'p)T,(y'0,y'1,...,y'p)T]。
Based on the SAVD, the displacement vector calculation is carried out on the block areas where the central points of all the block areas are located, and the displacement vector result Lx ═ of (Lx) of each area block in each cloud cluster area is output1,lx2,…,lxp),Ly＝(ly1,ly2,…,lxp)。
As shown in fig. 2, a complete flow of the cloud displacement vector calculation method based on cloud edge identification provided in this embodiment includes: on the basis of cloud image binaryzation, firstly determining an expansion parameter width to realize the region division of all cloud clusters in the space; then, determining the composition of pixel points on the outer edge of each cloud cluster area by adopting a cloud cluster outer edge searching method and based on an edge type judgment strategy; sequentially determining coordinates of the center points of the block areas along the outer edges of the cloud cluster by adopting a block area center positioning algorithm, and performing block area overlapping degree optimization and block area center point position correction; and finally, according to the obtained block region position, calling a block matching method SAVD based on the sum of absolute value differences to calculate the displacement vector of each block region, and further obtaining the block region displacement vector calculation result of each cloud cluster region. According to the embodiment, the displacement vector information of different positions of each cloud cluster can be expressed, and the calculation efficiency and the calculation precision are improved.
The cloud displacement vector calculation device based on cloud edge identification according to the embodiment of the present invention is described below, and the cloud displacement vector calculation device based on cloud edge identification described below and the cloud displacement vector calculation method based on cloud edge identification described above may be referred to in correspondence with each other.
As shown in fig. 3, the apparatus includes a segmentation module 301, an extraction module 302, a determination module 303, and a calculation module 304, wherein:
the segmentation module 301 is configured to segment a cloud cluster region from a current cloud atlas;
wherein, the current cloud picture is the cloud picture which needs to be subjected to displacement vector calculation. The cloud picture comprises a cloud cluster area and a background area. And segmenting the cloud cluster region from the current cloud picture by using a segmentation algorithm. The present embodiment is not limited to a specific segmentation method.
The extraction module 302 is used for extracting the outer edge of each cloud cluster area;
the segmentation result of the current cloud picture can be regarded as a binary cloud picture. Wherein, the pixel values of the background area in the binary cloud picture are the same and are 0; the pixel values of the cloud cluster regions are all the same and are 1. Therefore, the binary cloud picture cannot reflect the motion information of the cloud cluster, and only the edge motion information of the cloud cluster is valuable. Therefore, the present embodiment extracts the outer edge of each cloud region.
The determining module 303 is configured to determine, on an outer edge of each cloud cluster region, a central point of each block region distributed along the outer edge, and determine a position of each block region according to the central point of each block region;
after the outer edge of each cloud cluster area is obtained, an edge point is selected from the outer edge as the center point of the block area. The present embodiment is not limited to a specific selection method. The selection of the center point of the patch area causes the patch area to be evenly distributed on the outer edge of the cloud cluster area.
The calculating module 304 is configured to calculate a displacement vector of each block region based on a block matching algorithm according to the position of each block region.
The block matching algorithm is based on a partition area with a certain size in the current cloud picture, and an optimal matching partition area is searched in a certain range of frames before and after the current cloud picture according to a certain measurement method. The present embodiment does not limit the measurement method. And the displacement vector of each block area on the edge of each cloud image area is the displacement vector of the edge of each cloud image area.
In the embodiment, the outer edge of each cloud image region in the cloud image is extracted, the central point of each block region is determined on the outer edge of each cloud cluster region, and the displacement vector of each block region is calculated based on the block matching algorithm, so that on one hand, only the edge motion information of the cloud cluster is valuable for calculating the cloud image displacement vector, and therefore, the displacement vector is calculated only for each block region on the cloud cluster edge, and the calculation efficiency is improved; on the other hand, the displacement vectors at the positions on the outer edge of the cloud cluster are obtained through a block matching algorithm, so that the displacement vector information of different positions of the cloud cluster can be fully expressed, and the accuracy of cloud cluster displacement vector calculation is improved.
Fig. 4 illustrates a physical structure diagram of an electronic device, which may include, as shown in fig. 4: a processor (processor)410, a communication Interface 420, a memory (memory)430 and a communication bus 440, wherein the processor 410, the communication Interface 420 and the memory 430 are communicated with each other via the communication bus 440. The processor 410 may invoke logic instructions in the memory 430 to perform a cloud displacement vector calculation method based on cloud edge identification, the method comprising: dividing a cloud cluster area from the current cloud picture; extracting the outer edge of each cloud cluster region; determining the central point of each block region distributed along the outer edge on the outer edge of each cloud cluster region, and determining the position of each block region according to the central point of each block region; and calculating the displacement vector of each block area based on a block matching algorithm according to the position of each block area.
In addition, the logic instructions in the memory 430 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, an embodiment of the present invention further provides a computer program product, where the computer program product includes a computer program stored on a non-transitory computer-readable storage medium, where the computer program includes program instructions, and when the program instructions are executed by a computer, the computer is capable of executing the cloud displacement vector calculation method based on cloud edge identification provided in the above-mentioned method embodiments, where the method includes: dividing a cloud cluster area from the current cloud picture; extracting the outer edge of each cloud cluster region; determining the central point of each block region distributed along the outer edge on the outer edge of each cloud cluster region, and determining the position of each block region according to the central point of each block region; and calculating the displacement vector of each block area based on a block matching algorithm according to the position of each block area.
In yet another aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented by a processor to perform the cloud edge identification-based cloud displacement vector calculation method provided in the foregoing embodiments, the method including: dividing a cloud cluster area from the current cloud picture; extracting the outer edge of each cloud cluster region; determining the central point of each block region distributed along the outer edge on the outer edge of each cloud cluster region, and determining the position of each block region according to the central point of each block region; and calculating the displacement vector of each block area based on a block matching algorithm according to the position of each block area.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims (10)

1. A cloud cluster displacement vector calculation method based on cloud cluster edge identification is characterized by comprising the following steps:
dividing a cloud cluster area from the current cloud picture;
extracting the outer edge of each cloud cluster region;
determining the central point of each block region distributed along the outer edge on the outer edge of each cloud cluster region, and determining the position of each block region according to the central point of each block region;
and calculating the displacement vector of each block area based on a block matching algorithm according to the position of each block area.
2. The cloud cluster displacement vector computing method based on cloud cluster edge identification according to claim 1, wherein the segmenting of the cloud cluster region from the current cloud graph comprises:
segmenting the current cloud picture based on a threshold segmentation method to obtain an initial cloud cluster in the current cloud picture;
performing expansion operation on the initial cloud cluster based on the current expansion coefficient to generate a new cloud cluster;
counting the number of new clouds with the number of pixels larger than a preset pixel number threshold value, and if the number of the new clouds is smaller than the preset cloud number threshold value, increasing the current expansion coefficient;
performing a dilation operation on the initial cloud cluster again by using the increased current dilation coefficient until the number of the generated new cloud clusters is greater than or equal to the preset cloud cluster number threshold;
and taking the new cloud cluster generated in the last iteration as the cloud cluster area segmented from the current cloud picture.
3. The cloud displacement vector computing method based on cloud edge identification according to claim 1, wherein the extracting the outer edge of each cloud region comprises:
selecting any highest point from the outer edge of each cloud cluster area as an initial point for searching the outer edge of each cloud cluster area;
for the edge point of any cloud cluster area searched last time, searching a next edge point of the cloud cluster area according to the edge type of the edge point and the pixel point adjacent to the edge point, and determining the edge type of the next edge point until the next edge point is overlapped with the initial point of the outer edge of the cloud cluster area and the iteration number is more than 3;
wherein the edge types include an upper edge, a lower edge, a left edge, and a right edge; the upper side adjacent pixel point of the edge point with the edge type of the upper edge is the background of the current cloud picture, the lower side adjacent pixel point of the edge point with the edge type of the lower edge is the background, the left side adjacent pixel point of the edge point with the edge type of the left edge is the background, and the right side adjacent pixel point of the edge point with the edge type of the right edge is the background;
and taking all the found edge points of the cloud cluster area as pixel points of the outer edge of the cloud cluster area.
4. The cloud cluster displacement vector computing method based on cloud cluster edge recognition according to claim 3, wherein the finding a next edge point of the cloud cluster region according to the edge type of the edge point and the pixel points adjacent to the edge point and determining the edge type of the next edge point comprises:
if the edge type of the edge point is an upper edge and the adjacent pixel point at the upper right corner of the edge point is a pixel point in the cloud cluster region, taking the adjacent pixel point at the upper right corner of the edge point as the next edge point, and taking the edge type of the next edge point as a left edge point;
if the edge type of the edge point is an upper edge, an adjacent pixel point at the upper right corner of the edge point is the background, and an adjacent pixel point at the right side is a pixel point in the cloud cluster region, taking the adjacent pixel point at the right side of the edge point as the next edge point, and taking the edge type of the next edge point as the upper edge point;
if the edge type of the edge point is an upper edge, and the adjacent pixel point at the upper right corner and the adjacent pixel point at the right side of the edge point are the backgrounds, taking the edge point as the next edge point, and taking the edge type of the next edge point as a right edge point;
if the edge type of the edge point is a lower edge and the adjacent pixel point at the lower left corner of the edge point is a pixel point in the cloud cluster area, taking the adjacent pixel point at the lower left corner of the edge point as the lower edge point, and taking the edge type of the lower edge point as a right edge point;
if the edge type of the edge point is a lower edge, the adjacent pixel point at the lower left corner of the edge point is the background, and the adjacent pixel point at the left side is the pixel point in the cloud cluster area, taking the adjacent pixel point at the left side of the edge point as the next edge point, and taking the edge type of the next edge point as the lower edge point;
if the edge type of the edge point is a lower edge, and the adjacent pixel point at the lower left corner and the adjacent pixel point at the left side of the edge point are the backgrounds, taking the edge point as the lower edge point, and taking the edge type of the lower edge point as a left edge point;
if the edge type of the edge point is a left edge and the adjacent pixel point at the upper left corner of the edge point is a pixel point in the cloud cluster region, taking the adjacent pixel point at the upper left corner of the edge point as the next edge point, and taking the edge type of the next edge point as the lower edge point;
if the edge type of the edge point is a left edge, an adjacent pixel point at the upper left corner of the edge point is the background, and an adjacent pixel point at the upper side is a pixel point in the cloud cluster region, taking the adjacent pixel point at the left side of the edge point as the next edge point, and taking the edge type of the next edge point as the left edge point;
if the edge type of the edge point is a left edge, and the adjacent pixel point at the upper left corner and the adjacent pixel point at the upper side of the edge point are the backgrounds, taking the edge point as the next edge point, and taking the edge type of the next edge point as the upper edge point;
if the edge type of the edge point is the right edge and the adjacent pixel point at the lower right corner of the edge point is the pixel point in the cloud cluster area, taking the adjacent pixel point at the upper right corner of the edge point as the next edge point, and taking the edge type of the next edge point as the upper edge point;
if the edge type of the edge point is the right edge, the adjacent pixel point at the lower right corner of the edge point is the background, and the adjacent pixel point at the lower side is the pixel point in the cloud cluster area, taking the adjacent pixel point at the lower side of the edge point as the next edge point, and taking the edge type of the next edge point as the right edge point;
and if the edge type of the edge point is the right edge, and the adjacent pixel point at the lower right corner and the adjacent pixel point at the lower side of the edge point are the backgrounds, taking the edge point as the next edge point, wherein the edge type of the next edge point is the lower edge point.
5. The cloud displacement vector computing method based on cloud edge identification according to any one of claims 1 to 4, wherein the determining the central point of each cloud region on the outer edge of each cloud region along the outer edge comprises:
sequentially sequencing the pixel points on the outer edge of each cloud cluster area along a clockwise sequence;
for the central point determined from the sorting result last time, retrieving a pixel point set from the sorting result, wherein the difference between the abscissa and the abscissa of the central point determined last time is equal to the side length of the block area, or the difference between the ordinate and the ordinate of the central point determined last time is equal to the side length of the block area;
and selecting the pixel point which is positioned behind the last determined central point in the sequencing result from the pixel point set, and taking the pixel point which is closest to the last determined central point in the selected pixel points as the central point of the block region until the closest pixel point is the last pixel point in the sequencing result.
6. The cloud displacement vector computing method based on cloud edge identification according to claim 5, wherein the step of determining the central point of each cloud region distributed along the outer edge of each cloud region further comprises:
sequencing the central points according to the sequence determined by the central points;
sequentially selecting a central point from the sequencing results of the central points;
calculating the overlapping degree between the block region where the central point is located and the block region where the selected central point is located for any central point behind the selected central point;
and if the overlapping degree is greater than a preset overlapping degree threshold value, deleting the central point from the sequencing result of the central points.
7. The cloud displacement vector computing method based on cloud edge identification according to claim 2, wherein the step of determining the central point of each cloud region distributed along the outer edge of each cloud region further comprises:
modifying the position of the central point of each block area according to the current expansion coefficient of the last expansion operation on the initial cloud cluster and the edge type of the central point of each block area;
wherein the edge types include an upper edge, a lower edge, a left edge, and a right edge; the upper side adjacent pixel point with the edge type being the central point of the upper edge is the background of the current cloud picture, the lower side adjacent pixel point with the edge type being the central point of the lower edge is the background, the left side adjacent pixel point with the edge type being the central point of the left edge is the background, and the right side adjacent pixel point with the edge type being the central point of the right edge is the background.
8. A cloud displacement vector calculation apparatus based on cloud edge recognition, comprising:
the segmentation module is used for segmenting cloud cluster areas from the current cloud picture;
the extraction module is used for extracting the outer edge of each cloud cluster area;
the determining module is used for determining the central point of each block region distributed along the outer edge on the outer edge of each cloud cluster region and determining the position of each block region according to the central point of each block region;
and the calculation module is used for calculating the displacement vector of each block area based on a block matching algorithm according to the position of each block area.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the cloud edge identification-based cloud displacement vector calculation method according to any one of claims 1 to 7 when executing the program.
10. A non-transitory computer readable storage medium, on which a computer program is stored, the computer program, when being executed by a processor, implementing the steps of the cloud displacement vector calculation method based on cloud edge identification according to any one of claims 1 to 7.
CN202011218307.0A 2020-11-04 2020-11-04 Cloud cluster displacement vector calculation method and device based on cloud cluster edge recognition Active CN112419407B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011218307.0A CN112419407B (en) 2020-11-04 2020-11-04 Cloud cluster displacement vector calculation method and device based on cloud cluster edge recognition

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011218307.0A CN112419407B (en) 2020-11-04 2020-11-04 Cloud cluster displacement vector calculation method and device based on cloud cluster edge recognition

Publications (2)

Publication Number Publication Date
CN112419407A true CN112419407A (en) 2021-02-26
CN112419407B CN112419407B (en) 2024-06-11

Family

ID=74826922

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011218307.0A Active CN112419407B (en) 2020-11-04 2020-11-04 Cloud cluster displacement vector calculation method and device based on cloud cluster edge recognition

Country Status (1)

Country Link
CN (1) CN112419407B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113205536A (en) * 2021-05-28 2021-08-03 湖北工业大学 Method for drawing aggregate density cloud picture

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20130011609A (en) * 2011-07-22 2013-01-30 에스케이플래닛 주식회사 Apparatus and method for sampled pixel based motion estimation
CN104766347A (en) * 2015-04-29 2015-07-08 上海电气集团股份有限公司 Cloud cluster movement prediction method based on foundation cloud chart
CN105225252A (en) * 2015-09-21 2016-01-06 华北电力大学（保定） Particle clouds motion Forecasting Methodology
CN107563397A (en) * 2016-06-30 2018-01-09 中国电力科学研究院 Cloud cluster method for calculation motion vector in a kind of satellite cloud picture

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20130011609A (en) * 2011-07-22 2013-01-30 에스케이플래닛 주식회사 Apparatus and method for sampled pixel based motion estimation
CN104766347A (en) * 2015-04-29 2015-07-08 上海电气集团股份有限公司 Cloud cluster movement prediction method based on foundation cloud chart
CN105225252A (en) * 2015-09-21 2016-01-06 华北电力大学（保定） Particle clouds motion Forecasting Methodology
CN107563397A (en) * 2016-06-30 2018-01-09 中国电力科学研究院 Cloud cluster method for calculation motion vector in a kind of satellite cloud picture

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
胡俊 等: "图像序列运动检测算法的研究及其应用", 武汉大学学报(自然科学版), vol. 46, no. 05, pages 1 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113205536A (en) * 2021-05-28 2021-08-03 湖北工业大学 Method for drawing aggregate density cloud picture
CN113205536B (en) * 2021-05-28 2022-04-29 湖北工业大学 Method for drawing aggregate density cloud picture

Also Published As

Publication number Publication date
CN112419407B (en) 2024-06-11

Similar Documents

Publication Publication Date Title
CN109903331B (en) Convolutional neural network target detection method based on RGB-D camera
CN108416307B (en) Method, device and equipment for detecting pavement cracks of aerial images
CN109784333B (en) Three-dimensional target detection method and system based on point cloud weighted channel characteristics
Shen et al. Detection of stored-grain insects using deep learning
CN111640125B (en) Aerial photography graph building detection and segmentation method and device based on Mask R-CNN
CN111179307A (en) Visual target tracking method for full-volume integral and regression twin network structure
CN107633226B (en) Human body motion tracking feature processing method
CN109446889B (en) Object tracking method and device based on twin matching network
WO2020165557A1 (en) 3d face reconstruction system and method
CN110322453A (en) 3D point cloud semantic segmentation method based on position attention and auxiliary network
CN109102514A (en) A kind of image partition method, device, equipment and computer readable storage medium
CN107146219B (en) Image significance detection method based on manifold regularization support vector machine
CN113888461A (en) Method, system and equipment for detecting defects of hardware parts based on deep learning
CN104715251A (en) Salient object detection method based on histogram linear fitting
CN107169962A (en) The gray level image fast partition method of Kernel fuzzy clustering is constrained based on space density
CN109712149A (en) A kind of image partition method based on wavelet energy and fuzzy C-mean algorithm
CN113963009B (en) Local self-attention image processing method and system based on deformable block
CN110310263B (en) SAR image residential area detection method based on significance analysis and background prior
CN111507340A (en) Target point cloud data extraction method based on three-dimensional point cloud data
CN112419407A (en) Cloud cluster displacement vector calculation method and device based on cloud cluster edge identification
CN113627481A (en) Multi-model combined unmanned aerial vehicle garbage classification method for smart gardens
CN107358625B (en) SAR image change detection method based on SPP Net and region-of-interest detection
CN117115575A (en) Improved RPCA infrared small target detection method based on scale space theory
CN112884884A (en) Candidate region generation method and system
CN110276260B (en) Commodity detection method based on depth camera

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant