CN116403007A - Remote sensing image change detection method based on target vector - Google Patents
Remote sensing image change detection method based on target vector Download PDFInfo
- Publication number
- CN116403007A CN116403007A CN202310390010.XA CN202310390010A CN116403007A CN 116403007 A CN116403007 A CN 116403007A CN 202310390010 A CN202310390010 A CN 202310390010A CN 116403007 A CN116403007 A CN 116403007A
- Authority
- CN
- China
- Prior art keywords
- target
- remote sensing
- vectors
- change
- image
- 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
Links
- 239000013598 vector Substances 0.000 title claims abstract description 106
- 230000008859 change Effects 0.000 title claims abstract description 61
- 238000001514 detection method Methods 0.000 title claims abstract description 61
- 238000002372 labelling Methods 0.000 claims abstract description 5
- 238000000034 method Methods 0.000 claims description 15
- 230000011218 segmentation Effects 0.000 claims description 5
- 230000009466 transformation Effects 0.000 claims description 5
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 4
- 238000013527 convolutional neural network Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 230000001131 transforming effect Effects 0.000 claims description 2
- 230000008569 process Effects 0.000 description 6
- 238000013135 deep learning Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 2
- 230000002238 attenuated effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/62—Extraction of image or video features relating to a temporal dimension, e.g. time-based feature extraction; Pattern tracking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/766—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using regression, e.g. by projecting features on hyperplanes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
- G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
-
- 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
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a remote sensing image change detection method based on a target vector, which comprises the following steps: sample labeling is carried out on the remote sensing sequence images by using vectors; constructing a target vector detection model, and inputting the marked remote sensing sequence image into the target vector detection model for training; detecting all targets in the remote sensing images of the same region at different times by using the target vector detection model to obtain target vectors of different sets; and calculating the similarity distances of the target vectors in different sets by using a change similarity algorithm to obtain the change condition of the target. By implementing the scheme of the invention, high-precision matching and refined change detection before and after target change in the remote sensing image can be realized.
Description
Technical Field
The invention relates to the technical field of deep learning, in particular to a remote sensing image change detection method based on a target vector.
Background
The change of the ship targets before and after the remote sensing images in the same area and at different times is significant for military use and civil use. In military aspect, the change monitoring can timely obtain the change condition of the ship target in the concerned area, and the battlefield dynamic sensing is automatically carried out, so that the combat strategy can be changed more rapidly by means of the change of remote sensing information; in civilian aspects, the system has great help to monitor, distribute and manage the port ships, and can increase the efficiency of port management. However, because the remote sensing images do not always have time continuity, the time of a plurality of remote sensing images obtained in the same area often has certain difference, a certain algorithm is needed for comparing the targets concerned with the front-back change, and particularly, for the conditions of large number of targets, frequent space change and large number of similar targets in ship change monitoring, a special algorithm is needed to ensure the accuracy of change detection.
The change detection is performed by manpower at the earliest, time and effort are consumed, and then with the appearance of a deep learning technology, the change detection is automatically realized by utilizing an intelligent model gradually, but the traditional deep learning change detection method ignores the accurate matching of the front and back changes of the target.
Disclosure of Invention
In order to solve the technical problems in the prior art, the invention aims to provide a remote sensing image change detection method based on a target vector, which realizes high-precision matching before and after target changes of ships and the like in remote sensing images and completes fine change detection.
In order to achieve the above purpose, the technical scheme of the invention is as follows:
the embodiment of the invention provides a remote sensing image change detection method based on a target vector, which comprises the following steps:
sample labeling is carried out on the remote sensing sequence images by using vectors;
constructing a target vector detection model, and inputting the marked remote sensing sequence image into the target vector detection model for training;
detecting all targets in the remote sensing images of the same region at different times by using the target vector detection model to obtain target vectors of different sets;
and calculating the similarity distances of the target vectors in different sets by using a change similarity algorithm to obtain the change condition of the target.
According to an aspect of the embodiment of the present invention, the sample labeling of the remote sensing sequence image by using the vector includes:
marking the target information of all targets in the remote sensing sequence image sample one by utilizing a minimum external inclined rectangular frame;
and (3) independently fusing the target information of each target to obtain a corresponding target vector.
According to an aspect of the embodiment of the present invention, the target information includes: the target type, four corner coordinates and center point coordinates of the minimum circumscribed inclined rectangular frame of the target, the target width, the target length and the shooting time of the remote sensing image containing the target.
According to one aspect of an embodiment of the present invention, the object vector detection model is composed of a backbone network and two branch networks,
the backbone network is a convolutional neural network formed by ResNet101, and targets are detected;
the two branch networks comprise a central point regression branch network and a bounding box regression branch network, wherein the central point regression branch network is used for predicting the central point position of the detection target, and the bounding box regression branch network is used for predicting the bounding box and the category of the detection target.
According to an aspect of the embodiment of the present invention, the inputting the labeled remote sensing sequence image into the target vector detection model for training includes:
performing segmentation pretreatment on the marked remote sensing sequence image, and reserving an image slice containing a complete target;
transforming the whole image coordinate system into a slice coordinate system, and forming a training data set by all image slices containing the complete target and the corresponding changed target vectors;
and inputting the training data set into the target vector detection model to train, wherein the training optimization standard is that the overall loss of two branch networks of the target vector detection model is reduced to the minimum, and a trained model parameter file is obtained at the moment and is used as a subsequent target vector detection model file.
According to one aspect of the embodiment of the invention, after the target vectors of different sets are obtained, affine transformation is performed on the coordinates of the target vectors of different sets by using a uniform geographic coordinate system, so that all the target vectors are in the same coordinate system.
According to an aspect of the embodiment of the present invention, the calculating the similarity distance of different set of target vectors by using the variable similarity algorithm includes:
the different sets comprise set a comprising vectors { a, B, c, d. }, set B comprising vectors { a ', B ', c ', d. };
calculating the change similarity between the corresponding target vectors in the set A and the set B of each target by using a similarity distance formula;
sorting the change similarity, reserving a target vector group with minimum change similarity, and setting a threshold value;
if the change similarity is larger than the threshold value, judging that the target disappears in the subsequent remote sensing image, if unpaired vectors appear in the remote sensing image after the time, judging that a new target appears, and judging the change position of the target according to the minimum change similarity principle.
Compared with the prior art, the invention has the following beneficial effects:
according to the scheme provided by the embodiment of the invention, the characteristics of invariance of the length and the width of the targets, such as ship targets and the like, time limitation of movement of the targets and the like are considered in the process of data marking, and more characteristic information of each target can be obtained in the process of target detection by marking the target information capable of reflecting the characteristics, so that a target vector is formed by sufficient target characteristic information. Therefore, the similarity between the targets can be calculated, sequenced and screened through the target vectors during the change detection, the change condition of the targets before and after different time sequences is obtained, and finally the high-precision change detection and the high-precision matching of the targets before and after the change detection are realized.
Compared with the traditional method, the scheme of the embodiment of the invention detects the target change in the remote sensing images of different time sequences in the same space through the target vector, thereby not only greatly improving the accuracy of the change detection result, but also effectively knowing the transformation condition of each target.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments will be briefly described below. It is apparent that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained from these drawings without inventive effort for a person of ordinary skill in the art.
Fig. 1 and fig. 2 schematically show a flowchart of a remote sensing image change detection method based on a target vector according to an embodiment of the present invention;
fig. 3 schematically shows a network structure diagram of a target vector detection model according to an embodiment of the present invention.
Detailed Description
The description of the embodiments of this specification should be taken in conjunction with the accompanying drawings, which are a complete description of the embodiments. In the drawings, the shape or thickness of the embodiments may be enlarged and indicated simply or conveniently. Furthermore, portions of the structures in the drawings will be described in terms of separate descriptions, and it should be noted that elements not shown or described in the drawings are in a form known to those of ordinary skill in the art.
Any references to directions and orientations in the description of the embodiments herein are for convenience only and should not be construed as limiting the scope of the invention in any way. The following description of the preferred embodiments will refer to combinations of features, which may be present alone or in combination, and the invention is not particularly limited to the preferred embodiments. The scope of the invention is defined by the claims.
As shown in fig. 1 and fig. 2, the embodiment of the invention discloses a remote sensing image change detection method based on a target vector, which comprises the following main steps:
s110, using vectors to carry out sample labeling on the remote sensing sequence images;
s120, constructing a target vector detection model, and inputting the marked remote sensing sequence image into the target vector detection model for training;
s130, detecting all targets in remote sensing images of the same region at different times by using a target vector detection model to obtain target vectors of different sets;
and S140, calculating the similarity distances of the target vectors in different sets by using a change similarity algorithm to obtain the change condition of the target.
In an exemplary embodiment, the implementation process of using the vector to sample the remote sensing sequence image in step S110 includes:
and marking the target information of all targets in the remote sensing sequence image sample one by utilizing the minimum external inclined rectangular frame. The target information includes: target class C,Four corner coordinates (X 0 ,Y 0 ,X 1 ,Y 1 ,X 2 ,Y 2 ,X 3 ,Y 3 ) And the center point coordinate O, the target width W, the target length L and the shooting time T of the remote sensing image containing the target. When the target is a ship, in other embodiments, the center position P of the ship 'S bow, the center position S of the ship' S stern, the direction D of movement, and the like, are also noted.
And (3) independently fusing the target information of each target to obtain a corresponding target vector. For example, each target uses a unique target vector [ C, (X) 0 ,Y 0 ,X 1 ,Y 1 ,X 2 ,Y 2 ,X 3 ,Y 3 ),O,W,L,T]To characterize.
As shown in fig. 3, in an exemplary embodiment, the object vector detection model in step S120 is composed of a backbone network and two branch networks. Wherein, the backbone network is a convolutional neural network formed by ResNet101, and the target is detected. The two branch networks include a center point regression branch network for predicting a center point position of the detection target and a bounding box regression branch network for predicting a bounding box and a category of the detection target.
For the center point regression branch network, the center point O is a positive sample, and other points are negative samples, but because the center point deviation also belongs to the normal range, the point loss in the Gaussian distribution range of the center point is attenuated to a certain extent, and the following loss function is used:
where p represents a predicted value, N represents the number of targets, i represents the target number, α and β are super parameters, where α=2, β=4 are selected.
In a bounding box regression branch network, a bounding box is formed by the type of a target in a target vector, the position of a target center point, the positions of four corner points of a target minimum circumscribed inclined rectangular box and the length of the targetAnd a broad definition, the parameters of the bounding box are defined as b= [ C, O, (X) 0 ,Y 0 ,X 1 ,Y 1 ,X 2 ,Y 2 ,X 3 ,Y 3 ),W,L]The following loss functions were used:
wherein b k Representing the actual value of the bounding box, b' k Represents the predicted value of the bounding box, N represents the number of targets, and Smooth L1 The expression of (2) is:
in an exemplary embodiment, the specific implementation process of inputting the marked remote sensing sequence image into the target vector detection model for training includes:
and carrying out segmentation pretreatment on the marked remote sensing sequence image, and reserving an image slice containing the complete target. The size of the segmentation and the step length of the segmentation are determined according to the detected target size, such as ships and the like.
The whole image coordinate system is transformed into a slice coordinate system, the target vector is changed along with the whole image coordinate system, and all image slices containing the complete target and the target vector which is correspondingly changed form a training data set.
And inputting the training data set into the target vector detection model for training, wherein the training optimization standard is that the overall loss of two branch networks (a central point regression branch network and a bounding box regression branch network) of the target vector detection model is reduced to the minimum, and a trained model parameter file is obtained at the moment and is used as a subsequent target vector detection model file.
In an exemplary embodiment, in step S130, all targets in the remote sensing images of the same region at different times are detected by using the target vector detection model, so as to obtain target vectors of different sets. That is, all targets of the same region are at different times using the target vector detection modelAnd detecting the front-back change in the remote sensing image. For example, ship targets in two remote sensing images (e.g., image a and image B) with different times are detected, and the detection results are characterized by using target vectors, wherein the target vectors are [ C, (X) 0 ,Y 0 ,X 1 ,Y 1 ,X 2 ,Y 2 ,X 3 ,Y 3 ),O,W,L,T]Wherein C represents the class of the ship target, and the coordinates of four corner points of the minimum circumscribed rectangular frame of the ship are (X) 0 ,Y 0 ,X 1 ,Y 1 ,X 2 ,Y 2 ,X 3 ,Y 3 ) The coordinate of the central point of the minimum circumscribed rectangular frame is O, the width of the ship is W, the length of the ship is L, and the shooting time of the ship image is T. And placing the target vectors of all targets in the preceding remote sensing image A in time into a set A, and placing the target vectors of all targets in the following remote sensing image B in time into a set B.
After step S130, affine transformation is performed on the coordinates of the target vectors of different sets using a uniform geographic coordinate system, so that all the target vectors are in the same coordinate system. Specifically, the transformed coordinates were calculated using the following formula of affine transformation:
XGeo＝GeoTransform[0]+GeoTransform[1]*X+Y*GeoTransform[2]
YGeo＝GeoTransform[3]+GeoTransform[4]*X+Y*GeoTransform[5]
wherein, six dimensions of GeoTansform [ 0-5 ] are respectively: the horizontal coordinate of the upper left corner of the image relative to the geographic coordinate, the horizontal resolution, the rotation parameter, the vertical coordinate of the upper left corner of the image relative to the geographic coordinate, the rotation parameter and the vertical resolution.
In an exemplary embodiment, the specific implementation process of calculating the similarity distances of the different set of target vectors in step S140 by using the varying similarity algorithm includes:
constructing a variable similarity algorithm, calculating the similarity of target vectors corresponding to each target in different sets of different time sequences in the same space, namely, calculating the distance between vectors after overlapping each dimension of the target vectors by different weights, and comparing the distance with the distanceSpatial distance. Different sets (these sets are sets of target vectors) include set a and set B, (in other embodiments, other sets corresponding to other time-series remote sensing images may be included), set a includes vectors { a, B, C, d. }, set B includes vectors { a ', B ', C ', d. }, where vector a may be [ C, (X) 0 ,Y 0 ,X 1 ,Y 1 ,X 2 ,Y 2 ,X 3 ,Y 3 ),O,W,L,T]Vector a ' may be [ C ', (X ] ' 0 ,Y’ 0 ,X’ 1 ,Y’ 1 ,X’ 2 ,Y’ 2 ,X’ 3 ,Y’ 3 ),O’,W’,L’,T’]；
The following similarity distance formula is used to calculate the varying similarity between the corresponding object vectors in set a and set B for each object,
where k is a shooting error coefficient, k=2; m is a distance movement coefficient, m=0.1; the smaller the similarity D of the object vectors, the closer the two objects are. The change condition of the targets before and after different time can be obtained according to the obtained similarity distance between each target in the remote sensing image A and each target in the remote sensing image B.
Sorting the change similarity, reserving a target vector group with minimum change similarity, and setting a threshold value;
if the change similarity is greater than the threshold value, judging that the target disappears in the subsequent remote sensing image, if unpaired vectors appear in the remote sensing image positioned at the time later, judging that a new target appears, and judging the change position of the target according to the minimum change similarity principle, for example, the change position of the target from the remote sensing image A to the remote sensing image B.
According to the scheme provided by the embodiment of the invention, the characteristics that targets such as ship targets and the like have length and width invariance, movement of the targets is limited by time and the like are considered in data marking, more characteristic information of each target can be obtained in target detection by marking the target information capable of reflecting the characteristics, and a target vector is formed by sufficient target characteristic information. Therefore, the similarity between the targets can be calculated, sequenced and screened through the target vectors during the change detection, the change condition of the targets before and after different time sequences is obtained, and finally the high-precision change detection and the high-precision matching of the targets before and after the change detection are realized.
The sequence numbers of the steps related to the method of the present invention do not mean the sequence of the execution sequence of the method, and the execution sequence of the steps should be determined by the functions and the internal logic, and should not limit the implementation process of the embodiment of the present invention in any way.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather to enable any modification, equivalent replacement, improvement or the like to be made within the spirit and principles of the invention.
Claims (7)
1. A remote sensing image change detection method based on a target vector comprises the following steps:
sample labeling is carried out on the remote sensing sequence images by using vectors;
constructing a target vector detection model, and inputting the marked remote sensing sequence image into the target vector detection model for training;
detecting all targets in the remote sensing images of the same region at different times by using the target vector detection model to obtain target vectors of different sets;
and calculating the similarity distances of the target vectors in different sets by using a change similarity algorithm to obtain the change condition of the target.
2. The method of claim 1, wherein the using the vector to sample the remote sensing sequence image comprises:
marking the target information of all targets in the remote sensing sequence image sample one by utilizing a minimum external inclined rectangular frame;
and (3) independently fusing the target information of each target to obtain a corresponding target vector.
3. The method of claim 2, wherein the target information comprises: the target type, four corner coordinates and center point coordinates of the minimum circumscribed inclined rectangular frame of the target, the target width, the target length and the shooting time of the remote sensing image containing the target.
4. The method of claim 1, wherein the object vector detection model is comprised of a backbone network and two branch networks,
the backbone network is a convolutional neural network formed by ResNet101, and targets are detected;
the two branch networks comprise a central point regression branch network and a bounding box regression branch network, wherein the central point regression branch network is used for predicting the central point position of the detection target, and the bounding box regression branch network is used for predicting the bounding box and the category of the detection target.
5. The method of claim 1, wherein inputting the annotated remote sensing sequence image into the target vector detection model for training comprises:
performing segmentation pretreatment on the marked remote sensing sequence image, and reserving an image slice containing a complete target;
transforming the whole image coordinate system into a slice coordinate system, and forming a training data set by all image slices containing the complete target and the corresponding changed target vectors;
and inputting the training data set into the target vector detection model to train, wherein the training optimization standard is that the overall loss of two branch networks of the target vector detection model is reduced to the minimum, and a trained model parameter file is obtained at the moment and is used as a subsequent target vector detection model file.
6. The method according to claim 1, characterized in that after obtaining the different sets of object vectors, affine transformation is performed on the coordinates of the different sets of object vectors using a uniform geographical coordinate system, so that all object vectors are in the same coordinate system.
7. The method of claim 1, wherein calculating similarity distances for different sets of target vectors using a varying similarity algorithm comprises:
the different sets comprise set a comprising vectors { a, B, c, d. }, set B comprising vectors { a ', B ', c ', d. };
calculating the change similarity between the corresponding target vectors in the set A and the set B of each target by using a similarity distance formula;
sorting the change similarity, reserving a target vector group with minimum change similarity, and setting a threshold value;
if the change similarity is larger than the threshold value, judging that the target disappears in the subsequent remote sensing image, if unpaired vectors appear in the remote sensing image after the time, judging that a new target appears, and judging the change position of the target according to the minimum change similarity principle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310390010.XA CN116403007B (en) | 2023-04-12 | 2023-04-12 | Remote sensing image change detection method based on target vector |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310390010.XA CN116403007B (en) | 2023-04-12 | 2023-04-12 | Remote sensing image change detection method based on target vector |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116403007A true CN116403007A (en) | 2023-07-07 |
| CN116403007B CN116403007B (en) | 2023-12-19 |
Family
ID=87017627
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310390010.XA Active CN116403007B (en) | 2023-04-12 | 2023-04-12 | Remote sensing image change detection method based on target vector |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116403007B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117808807A (en) * | 2024-02-29 | 2024-04-02 | 中国人民解放军国防科技大学 | Optical satellite remote sensing image instance level change detection method |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102542551A (en) * | 2010-12-13 | 2012-07-04 | 北京师范大学 | Automatic change detection technology for floating ice at edges of polar ice sheets |
| CN108573276A (en) * | 2018-03-12 | 2018-09-25 | 浙江大学 | A kind of change detecting method based on high-resolution remote sensing image |
| CN111652181A (en) * | 2020-06-17 | 2020-09-11 | 腾讯科技（深圳）有限公司 | Target tracking method and device and electronic equipment |
| CN111783523A (en) * | 2020-05-19 | 2020-10-16 | 中国人民解放军93114部队 | Remote sensing image rotating target detection method |
| CN113658146A (en) * | 2021-08-20 | 2021-11-16 | 合肥合滨智能机器人有限公司 | Nodule grading method and device, electronic equipment and storage medium |
| CN114022459A (en) * | 2021-11-09 | 2022-02-08 | 国网冀北电力有限公司 | Multi-temporal satellite image-based super-pixel change detection method and system |
| CN114972191A (en) * | 2022-04-25 | 2022-08-30 | 航天宏图信息技术股份有限公司 | Method and device for detecting farmland change |
| US20220292278A1 (en) * | 2021-03-12 | 2022-09-15 | Lawrence Livermore National Security, Llc | Model-based image change quantification |
| CN115410059A (en) * | 2022-11-01 | 2022-11-29 | 山东锋士信息技术有限公司 | Remote sensing image part supervision change detection method and device based on contrast loss |
| WO2022252799A1 (en) * | 2021-06-04 | 2022-12-08 | 成都数之联科技股份有限公司 | Model training method, woodland change detection method, system, and apparatus, and medium |
-
2023
- 2023-04-12 CN CN202310390010.XA patent/CN116403007B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102542551A (en) * | 2010-12-13 | 2012-07-04 | 北京师范大学 | Automatic change detection technology for floating ice at edges of polar ice sheets |
| CN108573276A (en) * | 2018-03-12 | 2018-09-25 | 浙江大学 | A kind of change detecting method based on high-resolution remote sensing image |
| CN111783523A (en) * | 2020-05-19 | 2020-10-16 | 中国人民解放军93114部队 | Remote sensing image rotating target detection method |
| CN111652181A (en) * | 2020-06-17 | 2020-09-11 | 腾讯科技（深圳）有限公司 | Target tracking method and device and electronic equipment |
| US20220292278A1 (en) * | 2021-03-12 | 2022-09-15 | Lawrence Livermore National Security, Llc | Model-based image change quantification |
| WO2022252799A1 (en) * | 2021-06-04 | 2022-12-08 | 成都数之联科技股份有限公司 | Model training method, woodland change detection method, system, and apparatus, and medium |
| CN113658146A (en) * | 2021-08-20 | 2021-11-16 | 合肥合滨智能机器人有限公司 | Nodule grading method and device, electronic equipment and storage medium |
| CN114022459A (en) * | 2021-11-09 | 2022-02-08 | 国网冀北电力有限公司 | Multi-temporal satellite image-based super-pixel change detection method and system |
| CN114972191A (en) * | 2022-04-25 | 2022-08-30 | 航天宏图信息技术股份有限公司 | Method and device for detecting farmland change |
| CN115410059A (en) * | 2022-11-01 | 2022-11-29 | 山东锋士信息技术有限公司 | Remote sensing image part supervision change detection method and device based on contrast loss |
Non-Patent Citations (2)
| Title |
|---|
| PEIQI TANG等: "PSNet: change detection with prototype similarity", 《 THE VISUAL COMPUTER 》, pages 3541 - 3550 * |
| 徐永杰: "基于高分影像的城市新增建设用地变化检测研发", 《中国优秀博硕士论文全文数据库（硕士）工程科技II辑》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117808807A (en) * | 2024-02-29 | 2024-04-02 | 中国人民解放军国防科技大学 | Optical satellite remote sensing image instance level change detection method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116403007B (en) | 2023-12-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111222574B (en) | Ship and civil ship target detection and classification method based on multi-model decision-level fusion | |
| CN106127204B (en) | A kind of multi-direction meter reading Region detection algorithms of full convolutional neural networks | |
| CN109740665A (en) | Shielded image ship object detection method and system based on expertise constraint | |
| CN109613006A (en) | A kind of fabric defect detection method based on end-to-end neural network | |
| CN109977945A (en) | Localization method and system based on deep learning | |
| CN111368769B (en) | Ship multi-target detection method based on improved anchor point frame generation model | |
| CN111882586B (en) | Multi-actor target tracking method oriented to theater environment | |
| CN110647802A (en) | Remote sensing image ship target detection method based on deep learning | |
| CN111340855A (en) | Road moving target detection method based on track prediction | |
| CN113033303B (en) | SAR image rotation ship detection implementation method based on RCIoU loss | |
| CN116403007B (en) | Remote sensing image change detection method based on target vector | |
| CN109658442A (en) | Multi-object tracking method, device, equipment and computer readable storage medium | |
| CN110633727A (en) | Deep neural network ship target fine-grained identification method based on selective search | |
| Liu et al. | Dynamic RGB-D SLAM based on static probability and observation number | |
| CN115294452B (en) | Rotary SAR ship target detection method based on bidirectional characteristic pyramid network | |
| CN109102538A (en) | Method, the synoptic analysis method of piston ring land characteristic point and land features line are extracted using isopleth data | |
| Wang et al. | Accurate real-time ship target detection using Yolov4 | |
| CN112364881A (en) | Advanced sampling consistency image matching algorithm | |
| CN107644203A (en) | A kind of feature point detecting method of form adaptive classification | |
| CN114549909A (en) | Pseudo label remote sensing image scene classification method based on self-adaptive threshold | |
| Zou et al. | Maritime target detection of intelligent ship based on faster R-CNN | |
| CN117689995A (en) | Unknown spacecraft level detection method based on monocular image | |
| CN117079120A (en) | Target recognition model optimization method based on improved GA algorithm | |
| CN115909072A (en) | Improved YOLOv4 algorithm-based impact point water column detection method | |
| CN114463628A (en) | Deep learning remote sensing image ship target identification method based on threshold value constraint |
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 |