CN113689481A - Quick matching algorithm for homonymous cloud points based on medium-resolution images - Google Patents

Quick matching algorithm for homonymous cloud points based on medium-resolution images Download PDF

Info

Publication number
CN113689481A
CN113689481A CN202111006452.7A CN202111006452A CN113689481A CN 113689481 A CN113689481 A CN 113689481A CN 202111006452 A CN202111006452 A CN 202111006452A CN 113689481 A CN113689481 A CN 113689481A
Authority
CN
China
Prior art keywords
image
matching
cloud
point
correlation coefficient
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.)
Withdrawn
Application number
CN202111006452.7A
Other languages
Chinese (zh)
Inventor
何永健
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.)
Nanjing University of Information Science and Technology
Original Assignee
Nanjing University of Information Science and Technology
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 Nanjing University of Information Science and Technology filed Critical Nanjing University of Information Science and Technology
Priority to CN202111006452.7A priority Critical patent/CN113689481A/en
Publication of CN113689481A publication Critical patent/CN113689481A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
    • 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
    • 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
    • G06T2207/10032—Satellite or aerial image; Remote sensing
    • 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
    • G06T2207/20—Special algorithmic details
    • G06T2207/20228—Disparity calculation for image-based rendering

Landscapes

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

Abstract

The invention discloses a quick matching algorithm of cloud points with the same name based on a medium-resolution image, which comprises the following steps: set the left image coordinate variable as (x)1,y1) The right image coordinate variable is (x)2,y2) Respectively setting the three continuously observed remote sensing images as AN, AF and AA images according to the observation time, and carrying out cloud and snow detection on the AN image, the AF image and the AA image; taking any one of AN AN image, AN AF image and AN AA image as a left image, taking the other images as right images, matching the right images by using the left image, giving AN initial matching position and left-right parallax values of a cloud and snow area, and setting a correlation coefficient threshold and a step length step of a point to be matched of the left image; and predicting the position of the target point on the left image on the right image according to the left and right parallax values. The matching algorithm takes a correlation coefficient method with high matching speed and high accuracy as a main matching means, and provides a cloud and snow detection method aiming at the condition of low matching density caused by cloud movement and cloud shadow influence, so as to eliminate the influence of cloud shadow.

Description

Quick matching algorithm for homonymous cloud points based on medium-resolution images
Technical Field
The invention relates to a digital image data information extraction method, which is a quick matching algorithm of cloud points with the same name based on a medium-resolution image.
Background
In the earth's atmospheric energy balance, clouds have a particularly pronounced regulatory role and are an important factor in influencing weather changes. The cloud top height and the cloud moving speed are basic macro parameters of the cloud and important data of weather forecast, the accurate cloud top height can improve the monitoring and forecasting accuracy of strong convection weather and guarantee aviation flight safety, and the method has important significance for analyzing the cloud cluster structure and height of local disastrous weather, such as strong storms, tropical cyclones and the like, cloud cluster three-dimensional space simulation, movement monitoring and the like.
The NASA (American space agency) utilizes MISR (detection satellite) images of three angles to obtain the cloud top height and the cloud moving speed through a reprojection technology, an image matching method and a geometric method. Therefore, the geometric correction precision of the image and the matching precision of the homonymous cloud points are main factors for restricting the height of the cloud top and the estimation precision of the cloud moving speed. The matching of the homonymous cloud points of the three images is the basis of resolving the cloud top height and the cloud moving speed, and the matching precision directly influences the precision of the result.
In the cloud point matching and cloud detection of the same name, numerous scholars perform cloud detection and cloud point matching of the same name by using a dynamic threshold method, a Logistic regression model (a generalized linear regression analysis model), SIFT (scale invariant feature transform), a least square method and a correlation coefficient method in sequence, and a certain effect is achieved.
However, most of images for cloud detection are medium-resolution images, so that landforms and clouds cannot be expressed finely, and due to the fact that scanning angles are different, shadows of different clouds and high reflectivity of the clouds are generated due to movement and deformation of the clouds, matching accuracy of the cloud points with the same name is low, and particularly, research is almost not conducted on matching of the triple image pair with the cloud points with the same name.
Disclosure of Invention
The invention aims to provide a quick matching algorithm of homonymous cloud points based on a medium-resolution image, which takes a correlation coefficient method with high matching speed and high accuracy as a main matching means, and provides a cloud and snow detection method aiming at the condition of low matching density caused by cloud movement and cloud shadow influence, so that the influence of cloud shadow is eliminated, and meanwhile, the matching amount of ground points is reduced; aiming at areas with small reflectivity change, the cloud points with the same name are matched and encrypted in an image enhancement mode, and through experiments, the algorithm is high in matching speed, and the matching success rate and the coverage rate can reach 100%.
The purpose of the invention can be realized by the following technical scheme:
a quick matching algorithm of cloud points with the same name based on a medium-resolution image is disclosed, and the matching algorithm is as follows:
three continuously shot medium-resolution remote sensing images at different angles are respectively AN image (AN), AN image (AF) and AN image (AA) according to the time sequence, and the shot remote sensing images are subjected to cloud point matching with the same name.
The same-name cloud point matching algorithm comprises the following steps:
s1, setting the left image coordinate variable as (x)1,y1) The right image coordinate variable is (x)2,y2) And respectively carrying out cloud and snow detection on the AN image, the AF image and the AA image.
S2, carrying out correlation coefficient method matching on the cloud and snow areas
Any one of the AN image, the AF image and the AA image is taken as a left image, other images are taken as right images, the left image is used for matching the right image, the initial matching position and the left-right parallax value of the cloud-snow area are given, and the correlation coefficient threshold value and the step length step of the point to be matched of the left image are set.
And S3, predicting the position of the target point on the left image on the right image in S2 according to the left and right parallax values.
S4 finding the target point (x) on the left image in S21,y1) Taking a target window with odd length as a center, taking a search window on the right image by taking the target point position predicted in S3 as the center, carrying out experimental scaling according to the gray level characteristics and the matching effect of the image, removing the interference of the ground and the cloud image in the search window, taking the gray level value of a pixel on the ground and the cloud image as 0, and taking the value of the pixel on the cloud and snow as a normal value.
And S5, calculating the correlation coefficient of each candidate point in the right image search window according to the correlation coefficient formula.
And S6, converting the left image and the right image, changing the point to be matched of the left image into the predicted position of the new right image, changing the point successfully matched into the target point of the new left image, and performing reverse matching.
And S7, taking the points successfully matched in the reverse direction in the S6 as new matching points, calculating new left and right parallax values in the x and y directions, returning to the S3, and continuously matching the next target point according to step steps until all cloud and snow points to be matched are matched.
And S8, recording the regions which cannot be matched, enhancing the highlight region, and repeating the steps of 2-7 to realize matching encryption.
Further, in S2, a left image coordinate variable (x) is set1,y1) And right image coordinate variable (x)2,y2) The corresponding positions of the initial matching points are respectively (x)10,y10) And (x)20,y20)。
Setting left and right parallax values in x and y directions as dx＝x20-x10And dy＝y20-y10And the left and right parallax values are continuously updated along with the matching, multiple lines are adopted for matching with 1 pixel at intervals, and the target point positions on the matched left image are as follows:
(x1,y1)＝(x10+step,y10)。
further, the positions of the target points in S3 on the right image are: (x)2,y2)＝(x1+dx，y1+dy)。
Further, the relational equation in S5 is as follows:
Figure BDA0003237389180000031
wherein g is the gray scale value of the left image, g 'is the gray scale value of the right image, Sgg and Sg' g 'are the gray square sum of the left image and the right image, Sgg' is the sum of the gray products of the left image and the right image, and ρ is the peak value of the correlation coefficient of the candidate point.
If the peak value ρ of the correlation coefficient of the candidate point is smaller than the threshold value of the correlation coefficient in S2, skipping the candidate point, keeping the left and right disparity values unchanged, adding a step to continue the matching of the next point, returning to step S3, and if the peak value ρ of the correlation coefficient of the candidate point is larger than the threshold value of the correlation coefficient in S2, locating the corresponding right image point (x < x >) thereof2,y2) As a point where the forward matching is successful, then the reverse matching is performed.
Further, the reverse matching step in S6 is as follows:
and (4) taking a target window and a search window according to the S4, calculating a correlation coefficient according to the S5, and if the peak value rho of the correlation coefficient of the candidate point is greater than the threshold value of the correlation coefficient in the S2, taking the corresponding new right image point as a point with successful reverse matching and comparing the new right image point with a target point in forward matching.
If the correlation coefficient peak value rho of the candidate point is smaller than the correlation coefficient threshold value in the S2, or the correlation coefficient peak value rho of the candidate point is not the same as the target point in the forward matching, the matching is unsuccessful, the point is skipped, the left and right parallax values are unchanged, a step length is added, and the step S3 is returned to continue the matching of the next point.
Further, in S7, the new left and right disparity values in the x and y directions are: dx＝x2-x1And dy＝y2-y1。
Further, in S8, the highlight is enhanced by using an image enhancement algorithm.
The invention has the beneficial effects that:
1. the matching algorithm takes a correlation coefficient method with high matching speed and high accuracy as a main matching means, and provides a cloud (snow) detection method aiming at the condition of low matching density caused by cloud movement and cloud shadow influence, so that the influence of cloud shadow is eliminated, and meanwhile, the ground point matching amount is reduced;
2. the matching algorithm of the invention adopts an image enhancement mode to carry out matching encryption on cloud points with the same name aiming at areas with small reflectivity change, and through experiments, the matching speed of the algorithm is high, and the matching success rate and the coverage rate can reach 100%.
Drawings
The invention will be further described with reference to the accompanying drawings.
FIG. 1 is a flow chart of the matching algorithm of the present invention;
FIG. 2 is an AF ground artwork for the invention;
FIG. 3 is AN AN of the present invention without going to ground artwork;
FIG. 4 is a schematic illustration of the AA ground artwork of the present invention;
FIG. 5 is a graph of AF-to-ground piecewise linear enhancement of the present invention;
FIG. 6 is a graph of the AN to ground piecewise linear enhancement of the present invention;
FIG. 7 is a graph of the AA surface-to-ground piecewise linear enhancement of the present invention.
Detailed Description
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 only a part of the embodiments of the present invention, and not all of the embodiments. 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 quick matching algorithm for cloud points with the same name based on a medium-resolution image, as shown in fig. 1, the matching algorithm is as follows:
three continuously shot medium-resolution remote sensing images at different angles are respectively AN image (AN), AN image (AF) and AN image (AA) according to the time sequence, and the shot remote sensing images are subjected to cloud point matching with the same name.
The cloud point matching algorithm with the same name comprises the following steps:
s1, setting the left image coordinate variable as (x)1,y1) The right image coordinate variable is (x)2,y2) And respectively carrying out cloud and snow detection on the AN image, the AF image and the AA image.
S2, carrying out correlation coefficient method matching on the cloud and snow areas
Taking any one of the AN image, the AF image and the AA image as a left image, taking the other images as right images, matching the right image by the left image, setting AN initial matching position and a left-right parallax value of a cloud-snow area, setting a correlation coefficient threshold value, and setting a left image coordinate variable (x)1,y1) And right image coordinate variable (x)2,y2) The corresponding positions of the initial matching points are respectively (x)10,y10) And (x)20,y20)。
Setting left and right parallax values in x and y directions as dx＝x20-x10And dy＝y20-y10The left and right parallax values are continuously updated along with the matching, and the points to be matched of the left image are setStep, adopting multi-line interval 1 pixel matching, and setting the target point position on the matched left image as follows:
(x1,y1)＝(x10+step,y10)
s3, predicting the position of the target point on the left image on the right image in S2 according to the left and right disparity values: (x)2,y2)＝(x1+dx，y1+dy)。
S4 finding the target point (x) on the left image in S21,y1) Taking a target window with odd length as a center, taking a search window on the right image by taking the target point position predicted in S3 as the center, carrying out experimental scaling according to the gray level characteristics and the matching effect of the image, removing the interference of the ground and the cloud image in the search window, taking the gray level value of a pixel on the ground and the cloud image as 0, and taking the value of the pixel on the cloud and snow as a normal value.
S5, calculating the correlation coefficient of each candidate point in the right image search window according to a correlation coefficient formula, wherein the correlation coefficient formula is as follows:
Figure BDA0003237389180000061
wherein g is the gray scale value of the left image, g 'is the gray scale value of the right image, Sgg and Sg' g 'are the gray square sum of the left image and the right image, Sgg' is the sum of the gray products of the left image and the right image, and ρ is the peak value of the correlation coefficient of the candidate point.
If the peak value ρ of the correlation coefficient of the candidate point is smaller than the threshold value of the correlation coefficient in S2, skipping the candidate point, keeping the left and right disparity values unchanged, adding a step to continue the matching of the next point, returning to step S3, and if the peak value ρ of the correlation coefficient of the candidate point is larger than the threshold value of the correlation coefficient in S2, locating the corresponding right image point (x < x >) thereof2,y2) As a point where the forward matching is successful, then the reverse matching is performed.
And S6, converting the left image and the right image, changing the point to be matched of the original left image into the predicted position of the new right image, changing the point which is successfully matched into the target point of the new left image, taking a target window and a search window according to S4, calculating a correlation coefficient according to S5, and if the peak value rho of the correlation coefficient of the candidate point is greater than the threshold value of the correlation coefficient in S2, taking the corresponding new right image point as the point which is successfully matched in the reverse direction and comparing the point with the target point in the forward direction matching.
If the correlation coefficient peak value rho of the candidate point is smaller than the correlation coefficient threshold value in the S2, or the correlation coefficient peak value rho of the candidate point is not the same as the target point in the forward matching, the matching is unsuccessful, the point is skipped, the left and right parallax values are unchanged, a step length is added, and the step S3 is returned to continue the matching of the next point.
S7, taking the points successfully matched in the reverse direction in the S6 as new matching points, and calculating left and right disparity values in the x and y directions, wherein the new left and right disparity values are respectively as follows: dx＝x2-x1And dy＝y2-y1And returning to the step S3 to continue to match the next target point according to the step until all the cloud snow points to be matched are matched.
And S8, recording the regions which cannot be matched, enhancing the highlight region by adopting an image enhancement algorithm, and repeating the steps of 2-7 to realize matched encryption.
The cloud point matching algorithm is verified by using a specific example, and the verification steps are as follows:
s11: data selection
Selecting medium-resolution remote sensing images of three different angles continuously shot by TERRA-MISR (multi-angle imaging spectrum radiometer detection satellite) with 275m spatial resolution, wherein the shooting angle of the AN image is 0 degree, the shooting angle of the AF image is 26.1 degrees, the shooting angle of the AA image is-26.1 degrees, and the AN image is taken as a left image to respectively match the AF image and the AA image which are taken as right images.
S12: cloud ground detection, parameter setting and matching with cloud points with same name
The gray threshold values in the correlation coefficient threshold values of the AN image, the AF image and the AA image are respectively set to be 67, 65 and 86, when the gray window of the image takes a value, the condition that the element gray value is larger than the gray threshold value is regarded as a cloud and snow pixel, the radius of the target window is set to be 5, the cloud and snow pixel matching is carried out by 5 pixels, the specific matching parameters are shown in table 1, and the same-name cloud point matching is carried out on the AN image, the AF image and the AA image.
S13: precision evaluation
The matching results of the improved algorithm and the improved algorithm are shown in table 1, table 2 is the comparison analysis results of the matching speed, density and precision of the unmodified algorithm and the improved algorithm, the matching points of the three images when the same-name point is found are shown by round dots and triangular dots, as shown in fig. 2-7, the matching points with larger errors are marked by x characters, and as can be seen from fig. 2-7 and table 2, the matching time of the unmodified algorithm is more because the ground is not removed and enhanced.
The matching success rate is relatively low when the same cloud and snow points can not be obtained in the high-brightness cloud and snow, the matching speed, the matching coverage rate and the matching success rate are improved by the improved algorithm, and the matching coverage rate and the matching success rate are obviously improved particularly in the aspect of matching of key cloud and snow.
TABLE 1 match parameter settings
Figure BDA0003237389180000081
TABLE 2 comparison of speed, density, accuracy of unmodified and modified algorithm matching
Figure BDA0003237389180000091
In the description herein, references to the description of "one embodiment," "an example," "a specific example" or the like are intended to mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The foregoing shows and describes the general principles, essential features, and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed.

Claims (7)

1. A quick matching algorithm of cloud points with the same name based on a medium-resolution image is characterized in that the matching algorithm is as follows:
three continuously shot medium-resolution remote sensing images at different angles, wherein the three images are respectively AN image (AN), AN image (AF) and AN image (AA) according to the time sequence, and the shot remote sensing images are subjected to same-name cloud point matching;
the same-name cloud point matching algorithm comprises the following steps:
s1, setting the left image coordinate variable as (x)1,y1) The right image coordinate variable is (x)2,y2) Respectively carrying out cloud and snow detection on the AN image, the AF image and the AA image;
s2, carrying out correlation coefficient method matching on the cloud and snow areas
Taking any one of AN AN image, AN AF image and AN AA image as a left image, taking the other images as right images, matching the right images by using the left image, giving AN initial matching position and left-right parallax values of a cloud and snow area, and setting a correlation coefficient threshold and a step length step of a point to be matched of the left image;
s3, predicting the position of the target point on the left image on the right image in S2 according to the left and right parallax values;
s4 finding the target point (x) on the left image in S21,y1) Taking a target window with odd length as a center, taking a search window on the right image by taking the target point position predicted in S3 as the center, carrying out experimental scaling according to the gray level characteristics and the matching effect of the image, removing the interference of the ground and the cloud image in the search window, taking the gray level value of a pixel on the ground and the cloud image as 0, and taking the value of the pixel on the cloud and snow as a normal value;
s5, calculating the correlation coefficient of each candidate point in the right image search window according to a correlation coefficient formula;
s6, converting the left image and the right image, changing the point to be matched of the original left image into the predicted position of the new right image, changing the point successfully matched into the target point of the new left image, and performing reverse matching;
and S7, taking the points successfully matched in the reverse direction in the S6 as new matching points, calculating new left and right parallax values in the x and y directions, returning to the S3, and continuously matching the next target point according to step steps until all cloud and snow points to be matched are matched.
And S8, recording the regions which cannot be matched, enhancing the highlight region, and repeating the steps of 2-7 to realize matching encryption.
2. The algorithm for fast matching of cloud points with same name based on medium resolution ratio image as claimed in claim 1, wherein the left image coordinate variable (x) is set in S21,y1) And right image coordinate variable (x)2,y2) The corresponding positions of the initial matching points are respectively (x)10,y10) And (x)20,y20)；
Setting left and right parallax values in x and y directions as dx＝x20-x10And dy＝y20-y10And the left and right parallax values are continuously updated along with the matching, multiple lines are adopted for matching with 1 pixel at intervals, and the target point positions on the matched left image are as follows:
(x1,y1)＝(x10+step,y10)。
3. the algorithm of claim 2, wherein the positions of the target points in S3 on the right image are: (x)2,y2)＝(x1+dx，y1+dy)。
4. The algorithm for fast matching of cloud points with same names based on medium resolution ratio image as claimed in claim 3, wherein the relational equation in S5 is as follows:
Figure FDA0003237389170000021
wherein g is the gray value of the left image, g 'is the gray value of the right image, Sgg and Sg' g 'are the gray square sum of the left image and the right image respectively, Sgg' is the sum of the gray products of the left image and the right image, and ρ is the peak value of the correlation coefficient of the candidate point;
if the peak value ρ of the correlation coefficient of the candidate point is smaller than the threshold value of the correlation coefficient in S2, skipping the candidate point, keeping the left and right disparity values unchanged, adding a step to continue the matching of the next point, returning to step S3, and if the peak value ρ of the correlation coefficient of the candidate point is larger than the threshold value of the correlation coefficient in S2, locating the corresponding right image point (x < x >) thereof2,y2) As a point where the forward matching is successful, then the reverse matching is performed.
5. The algorithm for fast matching of cloud points with same names based on medium resolution ratio image as claimed in claim 4, wherein the inverse matching step in S6 is as follows:
and (4) taking a target window and a search window according to the S4, calculating a correlation coefficient according to the S5, and if the peak value rho of the correlation coefficient of the candidate point is greater than the threshold value of the correlation coefficient in the S2, taking the corresponding new right image point as a point with successful reverse matching and comparing the new right image point with a target point in forward matching.
If the correlation coefficient peak value rho of the candidate point is smaller than the correlation coefficient threshold value in the S2, or the correlation coefficient peak value rho of the candidate point is not the same as the target point in the forward matching, the matching is unsuccessful, the point is skipped, the left and right parallax values are unchanged, a step length is added, and the step S3 is returned to continue the matching of the next point.
6. The algorithm for fast matching of cloud points with same names based on medium resolution images as claimed in claim 1, wherein in S7, the new left and right disparity values in x and y directions are: dx＝x2-x1And dy＝y2-y1。
7. The algorithm for fast matching of cloud points with the same name based on medium resolution ratio images as claimed in any one of claims 1 to 6, wherein the manner of enhancing the highlight areas in S8 is to use an image enhancement algorithm.
CN202111006452.7A 2021-08-30 2021-08-30 Quick matching algorithm for homonymous cloud points based on medium-resolution images Withdrawn CN113689481A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111006452.7A CN113689481A (en) 2021-08-30 2021-08-30 Quick matching algorithm for homonymous cloud points based on medium-resolution images

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111006452.7A CN113689481A (en) 2021-08-30 2021-08-30 Quick matching algorithm for homonymous cloud points based on medium-resolution images

Publications (1)

Publication Number Publication Date
CN113689481A true CN113689481A (en) 2021-11-23

Family

ID=78584196

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111006452.7A Withdrawn CN113689481A (en) 2021-08-30 2021-08-30 Quick matching algorithm for homonymous cloud points based on medium-resolution images

Country Status (1)

Country Link
CN (1) CN113689481A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115615345A (en) * 2022-12-16 2023-01-17 中南大学 Ground surface deformation monitoring method based on photogrammetry color point cloud registration

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103489197A (en) * 2013-10-18 2014-01-01 武汉大学 Urban aerial image corner feature matching method

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103489197A (en) * 2013-10-18 2014-01-01 武汉大学 Urban aerial image corner feature matching method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
何永健: "《基于摄影测量的云顶高度和云移动速度耦合解算模型研究》", 《博士电子期刊》 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115615345A (en) * 2022-12-16 2023-01-17 中南大学 Ground surface deformation monitoring method based on photogrammetry color point cloud registration

Similar Documents

Publication Publication Date Title
CN107680120B (en) Infrared small target tracking method based on sparse representation and transfer limited particle filtering
CN104299244B (en) Obstacle detection method and device based on monocular camera
Frey et al. DEM-based SAR pixel-area estimation for enhanced geocoding refinement and radiometric normalization
CN109063549B (en) High-resolution aerial video moving target detection method based on deep neural network
CN103871039B (en) Generation method for difference chart in SAR (Synthetic Aperture Radar) image change detection
CN108805832B (en) Improved gray projection image stabilizing method suitable for tunnel environment characteristics
CN110827332A (en) Registration method of SAR image based on convolutional neural network
CN110796691B (en) Heterogeneous image registration method based on shape context and HOG characteristics
US11030478B1 (en) System and method for correspondence map determination
Jeong et al. Improved multiple matching method for observing glacier motion with repeat image feature tracking
CN108830808B (en) On-satellite infrared image stripe noise removing method based on similar line window mean value compensation
CN103198447A (en) Wind arrow field real-time measuring method based on satellite cloud pictures
WO2019215819A1 (en) Synthetic aperture radar image analysis system, synthetic aperture radar image analysis method, and synthetic aperture radar image analysis program
CN104079800A (en) Shaking preventing method for video image in video surveillance
CN101546429B (en) Detection method of small target of feature space
Ravanbakhsh et al. A comparative study of DEM registration approaches
KR101870084B1 (en) Method for Shadow Removal from High Resolution Satellite Imagery using Vertical Borders of Shadow Region
Wang et al. Highly accurate geometric correction for seriously oblique aero remote sensing image based on the piecewise polynomial model
CN113689481A (en) Quick matching algorithm for homonymous cloud points based on medium-resolution images
Parmehr et al. Automatic registration of optical imagery with 3d lidar data using local combined mutual information
CN116129280B (en) Method for detecting snow in remote sensing image
CN107451992B (en) Method and device for detecting SAR image change
CN115457022A (en) Three-dimensional deformation detection method based on real-scene three-dimensional model front-view image
Su et al. Hierarchical moving curved fitting filtering method based on LIDAR data
CN108564603B (en) SAR image change detection method based on improved Gaussian mixture model

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
WW01 Invention patent application withdrawn after publication

Application publication date: 20211123

WW01 Invention patent application withdrawn after publication