CN111461006A - Optical remote sensing image tower position detection method based on deep migration learning - Google Patents
Optical remote sensing image tower position detection method based on deep migration learning Download PDFInfo
- Publication number
- CN111461006A CN111461006A CN202010247180.9A CN202010247180A CN111461006A CN 111461006 A CN111461006 A CN 111461006A CN 202010247180 A CN202010247180 A CN 202010247180A CN 111461006 A CN111461006 A CN 111461006A
- Authority
- CN
- China
- Prior art keywords
- tower
- data
- remote sensing
- network
- sensing 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
-
- 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
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
-
- 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
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Molecular Biology (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Bioinformatics & Computational Biology (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
A method for detecting the position of an optical remote sensing image tower based on deep migration learning belongs to the technical field of image processing and comprises the following steps: the method comprises the following steps: collecting DOTA data and transmission line tower image data of a Pleiades satellite, and processing the collected data according to a standard VOC data format; step two: designing a deep neural network structure for tower detection, and providing a high-accuracy tower position detection neural network aiming at scenes such as different tower states, background types and illumination; step three: a deep migration learning method is introduced, the requirement of an algorithm on the detection data of the remote sensing image tower target is reduced, and model training is completed; step four: and testing and verifying the algorithm by using the remote sensing image containing the four tower states. The method is suitable for carrying out high-accuracy detection on the power transmission tower in the visible light remote sensing image under complex application scenes of various tower states, background types, illumination and the like, and provides support for safety monitoring of the power transmission line.
Description
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to an optical remote sensing image tower position detection method based on deep migration learning.
The high-voltage transmission tower is one of the most important infrastructure in the power transmission system, the operation state of the high-voltage transmission tower determines the operation stability and safety of the whole power grid, and the detection of the high-voltage transmission tower is the basis for monitoring the operation state of the high-voltage transmission tower and is also an important part for monitoring the high-voltage transmission line. On the other hand, with the rapid development of the optical remote sensing technology, the resolution of images acquired by satellites is higher and higher, and detailed information is richer and richer, so that the monitoring of the operation state of a power grid transmission line in a wide area by using the optical remote sensing images becomes the development direction of remote sensing application research. Generally, target information can be obtained from a remote sensing image by a visual interpretation method and a machine detection method, wherein the visual interpretation method is that a professional interpreter carries out manual extraction, identification, confirmation and detailed description on a target in the image, and the method is accurate, but consumes a large amount of manpower and material resources and is easily influenced by subjective states, experiences and the like of people; the other is machine detection, that is, target features in an image are extracted through a computer, and then the target in the image is detected in a supervision/unsupervised manner by using methods such as statistical pattern recognition, a neural network, a Boosting algorithm, a Support Vector Machine (SVM), an Artificial Immune System (Artificial Immune System) and the like.
At present, the detection and identification of the tower by using the optical image at home and abroad are mostly based on low-altitude visible light imaging (such as low-altitude unmanned imaging, vehicle-mounted camera imaging and the like). Under the low-altitude imaging condition, the target background is mostly the sky, and the background is comparatively clean, and present method mostly requires specific imaging visual angle in order to make things convenient for artifical extraction characteristic moreover, is difficult to direct application in remote sensing image shaft tower and detects.
In practice, situations such as various background types, variable imaging angles, weather changes and the like in imaging scenes of space-based or space-based platforms are complex, particularly, when a tower is naturally or artificially damaged, the state of the tower can be changed, the invariance characteristic process is extremely difficult to design manually, the method based on deep learning can realize that manual intervention is not needed in a characteristic extraction stage, and when a new mode occurs, a corresponding sample is only needed to be added to retrain the model, so that the model has stronger expandability. However, in order to ensure high accuracy of the detection classification model obtained by deep learning training, the following assumptions need to be satisfied: 1) the test sample and the training sample need to satisfy independent equal distribution conditions, namely an I.I.D. (independent and dispersive) hypothesis; 2) the number of training samples is sufficiently large. In a real machine learning application scenario, these two assumptions are difficult to satisfy.
The transfer learning refers to a learning process for applying knowledge learned in an old domain (or called a source domain) to a new domain (or called a target domain) by utilizing similarity between data, tasks or models, and the requirements of the transfer learning process on samples do not need to meet the two assumptions mentioned above. The transfer learning method reduces the risk of overfitting of the model by sharing knowledge between the source domain and the target domain, so that the detection recognition model can still be used under the condition of insufficient samples.
Disclosure of Invention
The invention provides an optical remote sensing image tower position detection method based on deep migration learning, aiming at the problems that in remote sensing image high-voltage transmission tower detection and recognition, the traditional method is complex in artificial characteristic design process and difficult to realize high-precision detection of the tower position in a remote sensing image, and the deep migration learning method is introduced for reducing the requirements of a deep learning algorithm on tower target data. The method is suitable for carrying out high-accuracy detection on the position of the transmission tower in the visible light remote sensing image under complex application scenes such as various tower states, background types and illumination, and provides support for safety monitoring of the transmission line.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
a method for detecting the position of an optical remote sensing image tower based on deep migration learning comprises the following steps:
the method comprises the following steps: acquiring original data of a tower and original data of a source domain of an optical remote sensing image, and marking and standardizing to obtain tower data and source domain data;
step two: dividing source domain data into a source domain training set, a source domain verification set and a source domain test set, and dividing tower original data into a target domain training set, a target domain verification set and a target domain test set;
step three: performing data enhancement on the source domain training set and the source domain verification set to obtain source domain training data for network training, and performing data enhancement on the target domain training set and the target domain verification set to obtain target domain training data for network training;
step four: constructing a deep neural network for tower target detection by superposing the Incep modules at the position 2 and adding an RPN module at the rear end of the network;
step five: training the deep neural network obtained in the fourth step by using the source domain training data obtained in the third step to obtain a detection network on the source domain;
step six: adjusting the learning rate to 1/30, and the number of network end classification categories to 2, and finely adjusting the detection network obtained in the fifth step by using the target domain training data obtained in the third step to obtain a detection model on the target domain;
step seven: and testing the detection model on the target domain obtained in the step six by using the target domain test set obtained in the step two.
Further, in the third step, data enhancement is respectively performed on the source domain training set, the source domain verification set, the target domain training set, and the target domain verification set according to any combination of the following enhancement methods:
1) in order to improve the robustness of the model to the imaging angle, the data of the rotation angle and the overturn are enhanced;
2) in order to improve the robustness of the model to illumination change, data enhancement for adjusting brightness and saturation is carried out;
3) in order to improve the robustness of the model to the state change of the power transmission tower, the data of the distortion is enhanced;
4) and in order to improve the robustness of the model to the size change of the power transmission tower, the zooming data is enhanced.
Further, in the first step, the source domain raw data is a DOTA data set.
Further, the fourth step comprises the following specific steps:
an inclusion module at 2 is superposed at the characteristic extraction stage of the front half part of the neural network, an RPN module is added at the rear end of the network, and the modules are connected through a convolutional layer or a pooling layer to construct a deep neural network for pole tower target detection.
Compared with the prior art, the invention has the beneficial effects that:
(1) the invention provides an optical remote sensing image tower position detection method based on depth migration learning, aiming at the problems that the existing method adopting low-altitude visible light imaging (such as low-altitude unmanned aerial vehicle imaging, vehicle-mounted camera imaging and the like) is difficult to be applied to optical remote sensing image high-accuracy tower position detection, the traditional optical characteristic manual construction process is complicated, the generalization capability is not strong and the like. The method is suitable for carrying out high-accuracy detection on the position of the power transmission tower in the visible light remote sensing image under complex application scenes of various tower states, background types, illumination and the like, and provides support for safety monitoring of the power transmission line.
(2) Aiming at the problems that image characteristics meeting the requirements on scale and angle, particularly the invariance and selectivity of the tower state are difficult to find in complex application scenes such as various tower states, background types and illumination in the position detection of the optical remote sensing image tower, the background of a transmission tower is usually complex and the like, the invention realizes the automatic and abstract expression of target characteristics under complex conditions by utilizing the functions of different network layers (a convolution layer, a pooling layer, a regression layer and the like) and combining with the functions of special network modules based on the deep learning theory and automatically completes the target detection. Compared with the traditional target detection process, the method can greatly improve the accuracy of tower target detection in complex application scenes such as various tower states, background types, illumination and the like.
(3) The invention aims to reduce the requirement of a tower target detection depth neural network on tower target data, realize tower target detection depth neural network training and introduce a depth migration learning method. The existing standard database is used as source domain data, a parameter migration mode is adopted, the detection model trained in a source domain is applied to a target (tower) domain, model parameters are modified in a small batch in a fine adjustment mode, the migration of tower detection knowledge is realized, and the method has important significance for the practical application of a transmission tower position detection method based on deep learning.
Drawings
FIG. 1 is a flow chart of a method for detecting the position of an optical remote sensing image tower based on deep transfer learning;
FIG. 2 is an example of a VOC standard data format;
FIG. 3 is a diagram of an inclusion feature extraction module;
FIG. 4 is a diagram of a RPN recommended region generation module;
FIG. 5 is a diagram of a deep neural network structure for tower target detection;
FIG. 6 is a diagram of samples in a DOTA database;
FIG. 7 is a diagram of a sample in a tower database;
fig. 8 shows the actual image test results.
Detailed Description
The technical solutions of the present invention are further described below with reference to the embodiments and the accompanying fig. 1 to 8, but not limited thereto, and all modifications or equivalent substitutions that do not depart from the spirit and scope of the technical solutions of the present invention should be covered by the protection scope of the present invention.
The invention aims at the problem that the manual design of a characteristic process is difficult in the position detection and identification of a remote sensing image high-voltage power transmission tower in the traditional method, adopts a deep learning method to construct a detection and identification model, and aims at the problem that a large number of high-quality labeled samples are difficult to obtain in the actual machine learning process, and takes the similarity between data, tasks or models into consideration in the migration and learning process, applies the knowledge learned in the old field (or called as a source field) to the new field (or called as a target field), introduces the migration and learning method, reduces the requirements of tower target detection and identification on tower sample data, and realizes end-to-end tower target high-accuracy detection and identification.
Detailed description of the invention
Referring to fig. 1, the present embodiment describes a method for detecting a position of an optical remote sensing image tower based on deep migration learning, the method includes the following steps:
the method comprises the following steps: acquiring DOTA data as source domain original data, and acquiring transmission line tower image data of a Pleiades satellite as target domain original data, and performing labeling, splitting and data enhancement processing on the source domain original data and the target domain original data;
step two: designing a deep neural network structure for tower detection, and providing a high-accuracy tower position detection neural network model aiming at scenes such as different tower states, background types, illumination and the like;
step three: a deep migration learning method is introduced, the requirements of the model on the image data of the power transmission line tower are reduced, and model training is completed;
step four: and testing and verifying the model by using the remote sensing image containing the four tower states.
Detailed description of the invention
In a specific embodiment, a method for detecting a position of an optical remote sensing image tower based on deep migration learning includes the following specific steps:
DOTA data is selected as source domain original data, transmission line tower image data of a visible light wave band of the Pleiades satellite is selected as target domain original data, the two parts of data are subjected to operations such as cutting and labeling to enable the data to be in accordance with a VOC data format, and the specific format after processing is shown in figure 2, so that the source domain data and the target domain data are obtained. Then dividing the source domain data into a source domain training set, a source domain verification set and a source domain test set according to the proportion of 8:1:1, dividing the target domain data into a target domain training set, a target domain verification set and a target domain test set according to the proportion of 8:1:1, and respectively performing data enhancement on the source domain training set, the source domain verification set, the target domain training set and the target domain verification set according to any combination of the following enhancement methods:
1) in order to improve the robustness of the model to the imaging angle, the data of the rotation angle and the turnover are enhanced, and the rotation in the range of 0-90 degrees and the vertical or horizontal turnover are carried out randomly;
2) in order to improve the robustness of the model to illumination change, data enhancement of brightness and saturation adjustment is carried out, and the brightness and the saturation are randomly adjusted to be 0.5-1.5 times of the original brightness and saturation;
3) in order to improve the robustness of the model to the state change of the power transmission tower and enhance the data of the distortion deformation, the proportion factor is adopted to randomly take 30-60, and the elastic coefficient is adopted to randomly take 4-7 of elastic deformation to perform the distortion deformation;
4) in order to improve the robustness of the model to the size change of the power transmission tower and enhance the zooming data, the random zooming image is 0.2-5 times of the original image.
And randomly combining each image in the training set by the enhancing method, wherein the quantity of the enhanced training sample data is about 5000.
Detailed description of the invention
In a specific embodiment, the method for detecting the position of the optical remote sensing image tower based on the deep migration learning specifically includes the following steps:
the neural network inclusion module is mainly characterized in that convolution kernels with different sizes are adopted in the transverse direction of the network, the convolution kernels with different sizes mean that features with different resolutions are extracted, and feature graphs extracted by convolution layers with different receptive fields are utilized for fusing features with different scales. In the actual calculation process, the convolution layer of the large-receptive-field convolution kernel in the inclusion structure is split into a plurality of small-convolution kernel convolution layers, so that the effects of reducing the number of parameters, increasing the nonlinear structure and increasing the network expression capacity are achieved.
The inclusion module is shown in fig. 3, where a) is a schematic structural diagram, and the module implements multi-scale feature extraction by using convolution kernels of different scales for the same input feature map, b) is a structure adopted in actual network computation, and splits a large-scale convolution kernel, and introduces a residual structure to reduce the risk of model degradation, and at the same time, performs depth adjustment on the feature map through a 1 × 1 convolution layer to facilitate fusion of features of different scales.
The regional recommendation network used in the present invention performs recommendation region determination after the network feature extraction structure (the first half of the network), and this type of network structure is also referred to as RPN structure. As shown in fig. 4, the RPN structure diagram is that the recommended region generating section removes regions that exceed the boundary of the original image, and maps recommended regions of different sizes to a feature map of a fixed size. The specific implementation process is that according to the size of the output required, the feature graph corresponding to the recommended region is partitioned according to the output size, then each block is subjected to maximum pooling, and finally the output with a fixed size is obtained.
The structure of the deep neural network for detecting the target of the optical remote sensing image tower is shown in FIG. 5.
The complexity of a model decision function is a necessary condition of overfitting, and a regularization item can be added to the loss function to reduce the overfitting risk by destroying the necessary condition, namely reducing the complexity of the decision function.
The final objective function is as follows:
detailed description of the invention
In a specific embodiment, the high-precision detection method for the high-resolution optical remote sensing image tower based on the deep migration learning includes the following specific steps:
the DOTA data set is generally used as large development and test data of a remote sensing image detection algorithm, and is partially shown in FIG. 6. compared with the original data of the target domain, the DOTA data set is remote sensing data, the distribution of bottom layer features is similar, the categories of the ground objects in the background are close, such as grassland, bare soil, water surface and the like in the background, and the categories of the bottom layer features are similar.
The detailed process of deep migration training of the whole network comprises the following steps:
1) independently training a network comprising a feature extraction part and an RPN part by using source domain data;
2) using the network in the step 1) to generate a recommended area and a corresponding label as a training sample, fixing the characteristic, extracting partial network parameters, and training the position and class prediction partial network. At the moment, a detection identification model of the network on the source domain is obtained;
3) on the basis of the source domain detection model obtained in the step 2), adjusting network training hyper-parameters (the learning rate is reduced to 1/30) and the classified category number (source domain data have a plurality of types of targets, and only one target domain is provided), and finely adjusting by using tower data to finally obtain the detection model on the target domain. The specific fine tuning process is as follows: and replacing the source domain data with the target domain data, and repeating the steps 1) and 2), wherein the model parameters of the step 1) are initialized to the detection model obtained in the step 2) in the source domain, and only the RPN network structure and the subsequent branch network are updated.
Detailed description of the invention
In a specific embodiment, the method for detecting the position of the optical remote sensing image tower based on the deep migration learning includes the following specific steps:
and (4) carrying out target detection by using the remote sensing image data containing four tower states and the detection model obtained in the third step, wherein the figure 8 is an actual image test result.
Claims (5)
1. A method for detecting the position of an optical remote sensing image tower based on deep migration learning is characterized by comprising the following steps: the method comprises the following steps:
the method comprises the following steps: acquiring original data of a tower and original data of a source domain of an optical remote sensing image, and marking and standardizing to obtain tower data and source domain data;
step two: dividing source domain data into a source domain training set, a source domain verification set and a source domain test set, and dividing tower data into a target domain training set, a target domain verification set and a target domain test set;
step three: performing data enhancement on the source domain training set and the source domain verification set to obtain source domain training data for network training, and performing data enhancement on the target domain training set and the target domain verification set to obtain target domain training data for network training;
step four: constructing a deep neural network for tower target detection by superposing the Incep modules at the position 2 and adding an RPN module at the rear end of the network;
step five: training the deep neural network obtained in the fourth step by using the source domain training data obtained in the third step to obtain a detection network on the source domain;
step six: adjusting the network training hyper-parameters and the network tail end classification category number, and finely adjusting the detection network obtained in the fifth step by using the target domain training data obtained in the third step to obtain a detection model on the target domain;
step seven: and testing the detection model on the target domain obtained in the step six by using the target domain test set obtained in the step two.
2. The optical remote sensing image tower position detection method based on deep migration learning of claim 1, wherein: in the third step, data enhancement is respectively carried out on the source domain training set, the source domain verification set, the target domain training set and the target domain verification set according to any combination of the following enhancement methods:
1) in order to improve the robustness of the model to the imaging angle, the data of the rotation angle and the overturn are enhanced;
2) in order to improve the robustness of the model to illumination change, data enhancement for adjusting brightness and saturation is carried out;
3) in order to improve the robustness of the model to the state change of the power transmission tower, the data of the distortion is enhanced;
4) and in order to improve the robustness of the model to the size change of the power transmission tower, the zooming data is enhanced.
3. The optical remote sensing image tower position detection method based on deep migration learning of claim 1, wherein: in the first step, the source domain raw data is a DOTA data set.
4. The optical remote sensing image tower position detection method based on deep migration learning of claim 1, wherein: the fourth step comprises the following specific steps:
an inclusion module at 2 is superposed at the characteristic extraction stage of the front half part of the neural network, an RPN module is added at the rear end of the network, and the modules are connected through a convolutional layer or a pooling layer to construct a deep neural network for pole tower target detection.
5. The optical remote sensing image tower position detection method based on deep migration learning of claim 1, wherein: and step six, fine adjustment process: and adjusting the learning rate of the network training hyper-parameters to 1/30 and the number of classification categories at the end of the network to be 2.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010247180.9A CN111461006B (en) | 2020-03-31 | 2020-03-31 | Optical remote sensing image tower position detection method based on deep migration learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010247180.9A CN111461006B (en) | 2020-03-31 | 2020-03-31 | Optical remote sensing image tower position detection method based on deep migration learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111461006A true CN111461006A (en) | 2020-07-28 |
CN111461006B CN111461006B (en) | 2023-07-25 |
Family
ID=71682424
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010247180.9A Active CN111461006B (en) | 2020-03-31 | 2020-03-31 | Optical remote sensing image tower position detection method based on deep migration learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111461006B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111898523A (en) * | 2020-07-29 | 2020-11-06 | 电子科技大学 | Remote sensing image special vehicle target detection method based on transfer learning |
CN112883840A (en) * | 2021-02-02 | 2021-06-01 | 中国人民公安大学 | Power transmission line extraction method based on key point detection |
CN113689399A (en) * | 2021-08-23 | 2021-11-23 | 长安大学 | Remote sensing image processing method and system for power grid identification |
CN117132903A (en) * | 2023-10-26 | 2023-11-28 | 江苏云幕智造科技有限公司 | Rotary satellite component identification method based on deep migration learning |
Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103218810A (en) * | 2013-03-27 | 2013-07-24 | 华北电力大学 | Semantic segmentation method for power tower/pole images |
CN204461366U (en) * | 2015-04-03 | 2015-07-08 | 中信戴卡股份有限公司 | A kind of axle cap groove depth on-line measuring device |
CN106971200A (en) * | 2017-03-13 | 2017-07-21 | 天津大学 | A kind of iconic memory degree Forecasting Methodology learnt based on adaptive-migration |
CN107451616A (en) * | 2017-08-01 | 2017-12-08 | 西安电子科技大学 | Multi-spectral remote sensing image terrain classification method based on the semi-supervised transfer learning of depth |
CN107657279A (en) * | 2017-09-26 | 2018-02-02 | 中国科学院大学 | A kind of remote sensing target detection method based on a small amount of sample |
US20180260607A1 (en) * | 2017-03-10 | 2018-09-13 | At&T Intellectual Property I, L.P. | Structure From Motion for Drone Videos |
CN108956640A (en) * | 2018-04-04 | 2018-12-07 | 山东鲁能智能技术有限公司 | Vehicle-mounted detection apparatus and detection method suitable for distribution line inspection |
CN109332928A (en) * | 2018-10-23 | 2019-02-15 | 江苏山扬智能装备有限公司 | Street lamp post robot welding system and welding method based on deep learning on-line checking |
CN109977918A (en) * | 2019-04-09 | 2019-07-05 | 华南理工大学 | A kind of target detection and localization optimization method adapted to based on unsupervised domain |
CN110210355A (en) * | 2019-05-24 | 2019-09-06 | 华南农业大学 | Weeds in paddy field category identification method and system, target position detection method and system |
CN110378252A (en) * | 2019-06-28 | 2019-10-25 | 浙江大学 | A kind of distress in concrete recognition methods based on depth migration study |
CN110780146A (en) * | 2019-12-10 | 2020-02-11 | 武汉大学 | Transformer fault identification and positioning diagnosis method based on multi-stage transfer learning |
US20200082221A1 (en) * | 2018-09-06 | 2020-03-12 | Nec Laboratories America, Inc. | Domain adaptation for instance detection and segmentation |
-
2020
- 2020-03-31 CN CN202010247180.9A patent/CN111461006B/en active Active
Patent Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103218810A (en) * | 2013-03-27 | 2013-07-24 | 华北电力大学 | Semantic segmentation method for power tower/pole images |
CN204461366U (en) * | 2015-04-03 | 2015-07-08 | 中信戴卡股份有限公司 | A kind of axle cap groove depth on-line measuring device |
US20180260607A1 (en) * | 2017-03-10 | 2018-09-13 | At&T Intellectual Property I, L.P. | Structure From Motion for Drone Videos |
CN106971200A (en) * | 2017-03-13 | 2017-07-21 | 天津大学 | A kind of iconic memory degree Forecasting Methodology learnt based on adaptive-migration |
CN107451616A (en) * | 2017-08-01 | 2017-12-08 | 西安电子科技大学 | Multi-spectral remote sensing image terrain classification method based on the semi-supervised transfer learning of depth |
CN107657279A (en) * | 2017-09-26 | 2018-02-02 | 中国科学院大学 | A kind of remote sensing target detection method based on a small amount of sample |
CN108956640A (en) * | 2018-04-04 | 2018-12-07 | 山东鲁能智能技术有限公司 | Vehicle-mounted detection apparatus and detection method suitable for distribution line inspection |
US20200082221A1 (en) * | 2018-09-06 | 2020-03-12 | Nec Laboratories America, Inc. | Domain adaptation for instance detection and segmentation |
CN109332928A (en) * | 2018-10-23 | 2019-02-15 | 江苏山扬智能装备有限公司 | Street lamp post robot welding system and welding method based on deep learning on-line checking |
CN109977918A (en) * | 2019-04-09 | 2019-07-05 | 华南理工大学 | A kind of target detection and localization optimization method adapted to based on unsupervised domain |
CN110210355A (en) * | 2019-05-24 | 2019-09-06 | 华南农业大学 | Weeds in paddy field category identification method and system, target position detection method and system |
CN110378252A (en) * | 2019-06-28 | 2019-10-25 | 浙江大学 | A kind of distress in concrete recognition methods based on depth migration study |
CN110780146A (en) * | 2019-12-10 | 2020-02-11 | 武汉大学 | Transformer fault identification and positioning diagnosis method based on multi-stage transfer learning |
Non-Patent Citations (1)
Title |
---|
刘小波等: "深度迁移学习在高光谱遥感图像分类中的研究现状与展望", 《青岛科技大学学报(自然科学版)》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111898523A (en) * | 2020-07-29 | 2020-11-06 | 电子科技大学 | Remote sensing image special vehicle target detection method based on transfer learning |
CN112883840A (en) * | 2021-02-02 | 2021-06-01 | 中国人民公安大学 | Power transmission line extraction method based on key point detection |
CN112883840B (en) * | 2021-02-02 | 2023-07-07 | 中国人民公安大学 | Power transmission line extraction method based on key point detection |
CN113689399A (en) * | 2021-08-23 | 2021-11-23 | 长安大学 | Remote sensing image processing method and system for power grid identification |
CN113689399B (en) * | 2021-08-23 | 2024-05-31 | 国网宁夏电力有限公司石嘴山供电公司 | Remote sensing image processing method and system for power grid identification |
CN117132903A (en) * | 2023-10-26 | 2023-11-28 | 江苏云幕智造科技有限公司 | Rotary satellite component identification method based on deep migration learning |
CN117132903B (en) * | 2023-10-26 | 2024-01-23 | 江苏云幕智造科技有限公司 | Rotary satellite component identification method based on deep migration learning |
Also Published As
Publication number | Publication date |
---|---|
CN111461006B (en) | 2023-07-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Henry et al. | Road segmentation in SAR satellite images with deep fully convolutional neural networks | |
CN111461006B (en) | Optical remote sensing image tower position detection method based on deep migration learning | |
WO2022111219A1 (en) | Domain adaptation device operation and maintenance system and method | |
CN114092832B (en) | High-resolution remote sensing image classification method based on parallel hybrid convolutional network | |
Tan et al. | Automatic extraction of built-up areas from panchromatic and multispectral remote sensing images using double-stream deep convolutional neural networks | |
CN110163069B (en) | Lane line detection method for driving assistance | |
CN110633708A (en) | Deep network significance detection method based on global model and local optimization | |
CN112950780B (en) | Intelligent network map generation method and system based on remote sensing image | |
CN104156734A (en) | Fully-autonomous on-line study method based on random fern classifier | |
CN112434723B (en) | Day/night image classification and object detection method based on attention network | |
CN109635726B (en) | Landslide identification method based on combination of symmetric deep network and multi-scale pooling | |
CN113838064B (en) | Cloud removal method based on branch GAN using multi-temporal remote sensing data | |
CN111882620A (en) | Road drivable area segmentation method based on multi-scale information | |
CN114863348B (en) | Video target segmentation method based on self-supervision | |
CN113591617B (en) | Deep learning-based water surface small target detection and classification method | |
CN113128620A (en) | Semi-supervised domain self-adaptive picture classification method based on hierarchical relationship | |
Chen et al. | Agricultural remote sensing image cultivated land extraction technology based on deep learning | |
CN112749675A (en) | Potato disease identification method based on convolutional neural network | |
CN113269224A (en) | Scene image classification method, system and storage medium | |
Liu et al. | Hierarchical multimodal fusion for ground-based cloud classification in weather station networks | |
CN114283285A (en) | Cross consistency self-training remote sensing image semantic segmentation network training method and device | |
CN114898097A (en) | Image recognition method and system | |
CN112949510A (en) | Human detection method based on fast R-CNN thermal infrared image | |
Wang et al. | Clouda: A ground-based cloud classification method with a convolutional neural network | |
CN116883650A (en) | Image-level weak supervision semantic segmentation method based on attention and local stitching |
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 |