CN112215085A - Power transmission corridor foreign matter detection method and system based on twin network - Google Patents
Power transmission corridor foreign matter detection method and system based on twin network Download PDFInfo
- Publication number
- CN112215085A CN112215085A CN202010983639.1A CN202010983639A CN112215085A CN 112215085 A CN112215085 A CN 112215085A CN 202010983639 A CN202010983639 A CN 202010983639A CN 112215085 A CN112215085 A CN 112215085A
- Authority
- CN
- China
- Prior art keywords
- power transmission
- transmission corridor
- image
- remote sensing
- change
- 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.)
- Pending
Links
- 230000005540 biological transmission Effects 0.000 title claims abstract description 60
- 238000001514 detection method Methods 0.000 title claims abstract description 48
- 230000008859 change Effects 0.000 claims abstract description 53
- 238000000034 method Methods 0.000 claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 9
- 238000002372 labelling Methods 0.000 claims abstract description 5
- 238000005070 sampling Methods 0.000 claims description 18
- 230000008569 process Effects 0.000 claims description 7
- 238000009432 framing Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 5
- 238000003384 imaging method Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 2
- 230000003595 spectral effect Effects 0.000 claims description 2
- 238000004364 calculation method Methods 0.000 abstract description 3
- 238000012423 maintenance Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 11
- 239000000203 mixture Substances 0.000 description 7
- 238000013528 artificial neural network Methods 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 230000011218 segmentation Effects 0.000 description 5
- 238000000605 extraction Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000012937 correction Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000003416 augmentation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001427 coherent effect Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
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
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Remote Sensing (AREA)
- Multimedia (AREA)
- Astronomy & Astrophysics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a twin network-based power transmission corridor foreign matter detection method and system, and belongs to the technical field of remote sensing. The method comprises the steps of firstly obtaining a multi-temporal remote sensing satellite image of a power transmission corridor, carrying out artificial fine labeling on a change target and a change region in the multi-temporal image, then slicing and carrying out binarization on the labeled region, then training a training set containing foreign matters in the power transmission corridor by using a twin network to obtain a model capable of detecting the foreign matters in the power transmission corridor, then detecting the multi-temporal remote sensing satellite image unknown whether the foreign matters in the power transmission corridor exist by using the obtained model, and finishing the judgment on whether the power transmission corridor contains the foreign matters. The method can detect and identify various foreign matters in the power transmission corridor, has the calculation efficiency which is obviously superior to that of a manual method, has higher foreign matter detection accuracy and reliability, and provides convenience for the operation and maintenance of the power transmission line.
Description
Technical Field
The invention belongs to the technical field of remote sensing, and particularly relates to a method and a system for detecting foreign matters in a power transmission corridor based on a twin network, which are used for detecting the change of a multi-temporal remote sensing satellite image in the power transmission corridor so as to find the foreign matters in the power transmission corridor.
Background
Foreign matter detection of the power transmission corridor relates to operations such as ground object identification and multi-temporal change analysis in a remote sensing image scene, and is closely connected with ground object target detection identification and change detection technology. The foreign matter detection method based on the ground object target recognition relies on a ground object extraction model with high enough recognition precision for a multi-temporal image scene to be detected, and on the basis of accurately extracting the positions and the contours of ground object targets of all scenes, targets which are likely to change are obtained by analyzing the changes of the appearances, the shapes and the positions of the targets; under the technical idea of foreign matter detection based on change detection, the algorithm model analyzes the change conditions of pixel values and image characteristics among multi-time-phase images by comparing remote sensing images of different time phases, and completes extraction and identification of foreign matters in a corridor based on identified difference areas among scenes.
For multi-temporal images for foreign matter detection, scene ground object types are complex in structure, the geographic area is wide, imaging elements such as weather and illumination among the multi-temporal images are obviously different, foreign matters in a power transmission corridor are diversified in size and types. Under the condition, the existing target detection model is difficult to ensure the detection and extraction precision of all surface feature targets in the scene, and the change detection model based on scene change analysis often introduces change noise information and cannot accurately position the change targets.
Therefore, how to overcome the defects of the prior art is a problem which needs to be solved urgently in the technical field of remote sensing at present.
Disclosure of Invention
The invention aims to solve the defects of the prior art and provides a method and a system for detecting foreign matters in a power transmission corridor based on a twin network.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
a power transmission corridor foreign matter detection method based on a twin network comprises the following steps:
step S1, acquiring multi-temporal remote sensing image data of a power transmission corridor area, selecting remote sensing images with imaging interval time more than one day and under the conditions of clear and cloudy weather, and selecting an image pair containing a foreign matter change target of the power transmission corridor;
step S2, constructing a training data set of the change detection model: preprocessing the remote sensing image obtained in the step S1, and then performing arrangement combination and symmetrical image data expansion on the remote sensing image;
step S3, carrying out object-level power transmission corridor change ground object labeling and region framing on the remote sensing image to be changed and detected;
step S4, slicing the pair of images marked in step S3 to generate a pair of changed image slices; then, binarizing the marked region to generate a binary mask change image pair;
step S5, a twin network model for change detection is constructed in a full supervision mode;
and step S6, detecting the newly acquired multi-temporal remote sensing image by using the twin network model constructed in the step S5, and detecting whether foreign matters exist in the power transmission corridor of the image.
Further, in step S2, the preprocessing includes geometric fine correction, registration, image mosaicing and cropping, cloud and shadow removal, and spectral normalization.
Further, in step S3, the GrabCut algorithm is preferably used for the power transmission corridor foreign matter profile-level framing method.
Further, in step S4, it is preferable that the image marked in step S3 is sliced into pieces of 572 × 572 pixels; and then, carrying out binarization on the manually marked change area by using a binarization method to generate a pool type binary mask change image pair, wherein the pair contains 1 foreign matters in the power transmission corridor and 0 foreign matters in the power transmission corridor.
Further, it is preferable that the specific method of step S5 is: when a twin network model for change detection is constructed, inputting a multi-temporal image slice and a change area matched with the multi-temporal image slice as the input of the network model, labeling a binarization mask, and outputting a judgment result of whether the slice area contains changes; the model main body adopts a Unet structure, the number of convolution kernels in an input layer of the Unet network is adjusted to be 3 according to the number of wave bands and the size of slices of an input remote sensing image, and the down-sampling level and the up-sampling level are respectively adjusted to be 4 levels.
Further, it is preferable that the specific method of step S6 is: and (4) carrying out change detection on the input remote sensing image pair by using the twin network model constructed in the step S5, respectively using the twin network to calculate the images of front and rear time phases, carrying out down-sampling and up-sampling by using Unet, fusing the feature map in the down-sampling process in the up-sampling process, and finally outputting a change mask, namely the power transmission corridor foreign matter mask.
The method utilizes the twin network to detect the change of the multi-temporal remote sensing satellite image of the power transmission corridor, and carries out the change detection with the resolution of 0.5 meter on common suspended objects such as kites, balloons, Kongming lantern and the like so as to realize the discovery of foreign matters, fully utilizes the difference analysis and identification capability of the image ground object types based on the deep learning model, and improves the accuracy of detecting and outputting the change information while improving the granularity and the detection sensitivity of a change induction space.
The method utilizes a twin network model based on a deep neural network to complete the detection and extraction of the variable foreign matters among multi-temporal images. Different from a common identification model, the twin network model can complete identification of feature change characteristics of land features related to semantics through fitting of manually set samples, and has the scene land feature identification capability and the change detection identification capability functionally. By means of the model, the foreign matter detection method can better complete scene foreign matter identification, and meanwhile, the detection and positioning accuracy of the foreign matters among multiple time intervals is improved by means of the change characteristic detection capability.
Compared with the prior art, the invention has the beneficial effects that:
the detection method can detect and identify various foreign matters in the power transmission corridor, has the calculation efficiency which is obviously superior to that of a manual method, has higher accuracy and reliability in foreign matter detection, and provides convenience for foreign matter detection in the power transmission corridor and safety maintenance of power transmission facilities in the corridor.
According to the method, by means of the semantic change recognition capability and the scene context analysis understanding comprehensive capability of the twin network model, the problem of target detection recognition error under a pure target detection technical scheme and the problem of low relevance of a target in a change area under the pure change detection technical scheme are effectively solved, and the detection recognition precision of the foreign matters in the power transmission corridor, which are of the same scale and uncertain type, is improved by fusing the ground object recognition information and the multi-temporal change information.
Drawings
FIG. 1 is a diagram of a Unet network architecture;
FIG. 2 is a diagram of a Siemens Network twin Network architecture;
FIG. 3 is a flow chart of a twin network based power transmission corridor foreign matter detection method.
Detailed Description
The present invention will be described in further detail with reference to examples.
It will be appreciated by those skilled in the art that the following examples are illustrative of the invention only and should not be taken as limiting the scope of the invention. The examples do not specify particular techniques or conditions, and are performed according to the techniques or conditions described in the literature in the art or according to the product specifications. The materials or equipment used are not indicated by manufacturers, and all are conventional products available by purchase.
A method for detecting foreign matters in a power transmission corridor based on a twin network is characterized in that a training set of the twin network for training multi-temporal remote sensing satellite images of an existing power transmission corridor is fully utilized to obtain a model with higher robustness and accuracy, and the obtained model is used for detecting the multi-temporal remote sensing satellite images of which the existence of the foreign matters in the power transmission corridor is unknown. The method for detecting the foreign matters in the power transmission corridor mainly comprises the following steps:
step S1, acquiring multi-temporal remote sensing image data of the power transmission corridor area through satellite shooting of the corresponding area of the power transmission corridor, wherein the format is a common remote sensing image format TIFF, selecting a remote sensing image under a proper time and a good weather condition, and selecting an image pair containing a foreign matter change target of the power transmission corridor;
step S2, constructing a training data set of the change detection model: after the remote sensing images are subjected to data preprocessing processes including geometric fine correction, registration, image mosaic and cutting, cloud removal, shadow processing and spectrum normalization, the data sets are subjected to permutation and combination and image data expansion in a symmetrical form, for example, t1and t2 … … tN time-varying images are subjected to permutation and combination, C (N,2) pairs of images are obtained, and A-B, B-A symmetrical augmentation can be performed on the image pairs after the permutation and combination;
step S3, marking and area framing of the object-level power transmission corridor variable ground object: the framing method for the foreign body profile level of the power transmission corridor uses GrabCut algorithm, and the GrabCut algorithm can perform fine profile generation of a target level. Implementation of the GrabCut algorithm:
and for an image containing N pixel points, GrabCT maps the image into a weighting network graph and establishes a corresponding energy function, and finally, the image is segmented through maximum flow/minimum segmentation. The segmentation results corresponding to the pixels in the picture are only of two types: one belonging to the foreground, denoted by 1, and one belonging to the background, denoted by 0. The GrabCont segmentation algorithm models the foreground and background in the image with a K-dimensional Gaussian Mixture Model (GMM), respectively. Wherein the probability density function of the single gaussian model is:
in the formula, x represents a pixel value of an image to be segmented, μ represents a pixel mean value of a local region of the image, σ represents a standard deviation of the pixel value of the local region of the image, and the gaussian mixture model is composed of n single gaussian models, as shown in the following formula:
it can be seen that the entire gaussian mixture model is the sum of the density weights of the single gaussian models, where t (i) is the weight of each gaussian model, and p (x) represents the probability value of the foreground at x pixel positions estimated using the mixture probability.
The GrabCut algorithm solves the optimal segmentation through an energy function. The energy function mainly includes two terms: a region penalty term and a boundary penalty term. The region penalty term is also called a data penalty term, and the boundary penalty term is mainly used for the relation between the space adjacent to the pixel point and the color. The energy function is represented in the form:
E(α,k,θ,z)＝U(α,k,θ,z)+V(α,z)
where E is the Gibbs energy, U is the data term, and V is the smoothing term. z represents the pixel value of the image, α represents the opacity, k represents the model parameter of a certain gaussian component, θ represents the coefficient of the gray histogram describing the gray level distribution composition of the foreground and background of the image, defined as θ ═ { h (z; α), α ═ 0,1}, and the data item U is defined as:
wherein D is a region term, n represents the number of Gaussian models for carrying out numerical fitting of data items, and is defined as:
D(αn,kn,θ,zn)＝-logp(zn|αn,kn,θ)-logπ(αn,kn)
and p (-) is a Gaussian probability distribution and π (-) is the mixture weight coefficient.
The gaussian parameter model at this time is:
θ＝{π(α,k),μ(α,k),∑(α,k),α＝0,1；k＝1…K}
the weight pi, the mean mu and the covariance sigma smoothing term V which respectively correspond to the kth Gaussian model can be solved by Euclidean distance of an RGB space:
in the above formula, γ is an adaptive parameter, β is a constant term, and C represents a pixel point sequence number set to be subjected to image segmentation. In the GrabCont algorithm, pixel points of an image are divided into 4 categories: { F, B, PF, PB }. Wherein F represents a determined foreground pixel point, B represents a determined background point, PF represents a possible foreground point, and PB represents a possible background point. In the GrabCont algorithm, a user selects an area needing to be segmented on an image by using a rectangular frame through manual marking, pixels outside the area are determined background pixels, and pixels which may be foreground or background in the area are collectively referred to as uncertainty items, so that a GMM (Gaussian mixture model) is established and initialized. Secondly, the GrabCut algorithm provides another interface for a user, and the user can specify which classes of { F, B, PF, PB } some pixels in the image belong to respectively, so that the accuracy of algorithm segmentation can be improved in a manual specifying mode.
Step S4, generating a pair of changed image slices and a mask for the labeled image: and (3) slicing the marked image pair, wherein the size of the slice is 572 pixel by 572 pixel, carrying out binarization on the originally marked area by using a binarization method, and generating a binary mask change image pair of the pool type, wherein the foreign matter in the power transmission corridor is 1, and the foreign matter in the power transmission corridor is not 0.
Step S5, constructing a twin network model for change detection: the twin network is a parallel network structure, and by means of the semantic level feature coding capability of the deep convolutional network, feature maps are simultaneously extracted from two different time phase images needing to be compared, and difference detection and positioning between the semantic level multi-time phase image pairs are realized through spatial difference analysis between the feature maps. In order to meet the multi-resolution target and area detection requirements of remote sensing images and realize multi-scale multi-temporal power transmission corridor change information positioning and identification, the Unet is used as a main Network structure of the Simase Network.
As in fig. 1, the network structure can be seen as 3 parts:
down-sampling: the dark downward arrow portion of the network represents pooling.
And (3) upsampling: the dark upward arrow portion of the network represents the deconvolution operation.
Softmax of the last layer: and outputting the characteristic diagram.
The unet network can be seen as firstly sampling and then convolving in different degrees, so that deep features are learned, the size of an original image is restored through upsampling, the upsampling is realized through deconvolution, and finally a feature map of the category number is output.
(1) Network architecture of the Siamese Network:
the sieme Network has two sub-networks (Input1and Input2) with the same structure and sharing weight, and two Input feeds are entered into two neural networks (Network 1and Network2), and the two neural networks respectively map the inputs to a new space to form a representation of the inputs in the new space. The 'disjunctive' of the neural network is realized by sharing weight values, and the similarity of two inputs can be evaluated through Loss calculation.
(2) Loss function for Siamese Network:
in a twin neural network (diameter network), the loss function adopted is contrast loss (contrast loss), and the relationship of paired data (paired data) in the twin neural network can be effectively processed by the loss function. The expression for the comparative loss is as follows:
wherein
The euclidean distance (two-norm) P representing the features X1 and X2 of the two samples represents the feature dimension of the samples, Y is a label indicating whether the two samples match, Y ═ 1 represents that the two samples are similar or match, Y ═ 0 represents mismatch, m is a set threshold, and N is the number of samples.
When Y is 1 (i.e., the samples are similar), the loss function is
That is, when the samples are not similar, if the euclidean distance of the feature space is small, the loss value becomes large.
When Y is 0 (i.e., the samples are not similar), the loss function is
That is, when the samples are not similar, if the euclidean distance of the feature space is small, the loss value becomes large.
Observing the expression of the coherent loss, the loss function can well express the matching degree of the pair samples, and can also be well used for training a model for extracting features.
Step S6, using the twin network model to detect the change of the multi-temporal remote sensing image: the method comprises the steps of carrying out change detection on an input remote sensing image by using a model obtained after training, calculating the input image through a twin network, carrying out down-sampling and up-sampling by using the Unet, fusing a feature map in the down-sampling process in the up-sampling process, and finally outputting a change mask, namely a power transmission corridor foreign matter mask, so as to detect whether a foreign matter exists in a power transmission corridor of the power transmission corridor.
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. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (6)
1. A power transmission corridor foreign matter detection method based on a twin network is characterized by comprising the following steps:
step S1, acquiring multi-temporal remote sensing image data of a power transmission corridor area, selecting remote sensing images with imaging interval time more than one day and under the conditions of clear and cloudy weather, and selecting an image pair containing a foreign matter change target of the power transmission corridor;
step S2, constructing a training data set of the change detection model: preprocessing the remote sensing image obtained in the step S1, and then performing arrangement combination and symmetrical image data expansion on the remote sensing image;
step S3, carrying out object-level power transmission corridor change ground object labeling and region framing on the remote sensing image to be changed and detected;
step S4, slicing the pair of images marked in step S3 to generate a pair of changed image slices; then, binarizing the marked region to generate a binary mask change image pair;
step S5, a twin network model for change detection is constructed in a full supervision mode;
and step S6, detecting the newly acquired multi-temporal remote sensing image by using the twin network model constructed in the step S5, and detecting whether foreign matters exist in the power transmission corridor of the image.
2. The twin network based power transmission corridor foreign object detection method according to claim 1, wherein in step S2, the preprocessing includes geometric refinement, registration, image mosaicing and cropping, cloud and shadow removal processing, and spectral normalization.
3. The twin network-based power transmission corridor foreign matter detection method according to claim 1, wherein in step S3, GrabCut algorithm is used for a power transmission corridor foreign matter profile-level framing method.
4. The method for detecting foreign objects in a power transmission corridor based on a twin network as claimed in claim 1, wherein in step S4, the image marked in step S3 is sliced into pieces of 572 by 572 pixels; and then, carrying out binarization on the manually marked change area by using a binarization method to generate a pool type binary mask change image pair, wherein the pair contains 1 foreign matters in the power transmission corridor and 0 foreign matters in the power transmission corridor.
5. The twin network-based power transmission corridor foreign matter detection method according to claim 1, wherein the specific method of step S5 is as follows: when a twin network model for change detection is constructed, inputting a multi-temporal image slice and a change area matched with the multi-temporal image slice as the input of the network model, labeling a binarization mask, and outputting a judgment result of whether the slice area contains changes; the model main body adopts a Unet structure, the number of convolution kernels in an input layer of the Unet network is adjusted to be 3 according to the number of wave bands and the size of slices of an input remote sensing image, and the down-sampling level and the up-sampling level are respectively adjusted to be 4 levels.
6. The twin network-based power transmission corridor foreign matter detection method according to claim 1, wherein the specific method of step S6 is as follows: and (4) carrying out change detection on the input remote sensing image pair by using the twin network model constructed in the step S5, respectively using the twin network to calculate the images of front and rear time phases, carrying out down-sampling and up-sampling by using Unet, fusing the feature map in the down-sampling process in the up-sampling process, and finally outputting a change mask, namely the power transmission corridor foreign matter mask.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010983639.1A CN112215085A (en) | 2020-09-17 | 2020-09-17 | Power transmission corridor foreign matter detection method and system based on twin network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010983639.1A CN112215085A (en) | 2020-09-17 | 2020-09-17 | Power transmission corridor foreign matter detection method and system based on twin network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112215085A true CN112215085A (en) | 2021-01-12 |
Family
ID=74050059
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010983639.1A Pending CN112215085A (en) | 2020-09-17 | 2020-09-17 | Power transmission corridor foreign matter detection method and system based on twin network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112215085A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113077424A (en) * | 2021-03-23 | 2021-07-06 | 广东电网有限责任公司广州供电局 | Power transmission line channel environment change detection method and system based on deep learning |
| CN113506316A (en) * | 2021-05-27 | 2021-10-15 | 北京迈格威科技有限公司 | Method and device for segmenting video object and network model training method |
| CN113743298A (en) * | 2021-09-03 | 2021-12-03 | 云南电网有限责任公司电力科学研究院 | Power grid foreign matter detection method and device based on satellite image deep learning |
| CN113822867A (en) * | 2021-09-24 | 2021-12-21 | 东莞市诺丽电子科技有限公司 | Train appearance abnormity detection method and train appearance abnormity detection system |
| CN115330794A (en) * | 2022-10-13 | 2022-11-11 | 扬州中科半导体照明有限公司 | LED backlight foreign matter defect detection method based on computer vision |
| CN115620150A (en) * | 2022-12-05 | 2023-01-17 | 海豚乐智科技(成都)有限责任公司 | Multi-modal image ground building identification method and device based on twin transform |
| CN118155082A (en) * | 2024-05-13 | 2024-06-07 | 山东锋士信息技术有限公司 | Double-branch change detection method based on hyperspectral space and spectral information |
| CN118410724A (en) * | 2024-07-02 | 2024-07-30 | 四川思极科技有限公司 | Transmission line foreign matter identification method, system, computer equipment and medium |
| CN118552800A (en) * | 2024-07-30 | 2024-08-27 | 广州英码信息科技有限公司 | Rail foreign matter semi-supervised anomaly detection method and system deployed at edge end |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN109409263A (en) * | 2018-10-12 | 2019-03-01 | 武汉大学 | A kind of remote sensing image city feature variation detection method based on Siamese convolutional network |
| CN111161218A (en) * | 2019-12-10 | 2020-05-15 | 核工业北京地质研究院 | High-resolution remote sensing image change detection method based on twin convolutional neural network |
| CN111259853A (en) * | 2020-02-04 | 2020-06-09 | 中国科学院计算技术研究所 | High-resolution remote sensing image change detection method, system and device |
| CN111640159A (en) * | 2020-05-11 | 2020-09-08 | 武汉大学 | Remote sensing image change detection method based on twin convolutional neural network |
| CN113298056A (en) * | 2021-07-27 | 2021-08-24 | 自然资源部国土卫星遥感应用中心 | Multi-mode remote sensing image change detection method, model generation method and terminal equipment |
| CN114022793A (en) * | 2021-10-28 | 2022-02-08 | 天津大学 | Optical remote sensing image change detection method based on twin network |
-
2020
- 2020-09-17 CN CN202010983639.1A patent/CN112215085A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN109409263A (en) * | 2018-10-12 | 2019-03-01 | 武汉大学 | A kind of remote sensing image city feature variation detection method based on Siamese convolutional network |
| CN111161218A (en) * | 2019-12-10 | 2020-05-15 | 核工业北京地质研究院 | High-resolution remote sensing image change detection method based on twin convolutional neural network |
| CN111259853A (en) * | 2020-02-04 | 2020-06-09 | 中国科学院计算技术研究所 | High-resolution remote sensing image change detection method, system and device |
| CN111640159A (en) * | 2020-05-11 | 2020-09-08 | 武汉大学 | Remote sensing image change detection method based on twin convolutional neural network |
| CN113298056A (en) * | 2021-07-27 | 2021-08-24 | 自然资源部国土卫星遥感应用中心 | Multi-mode remote sensing image change detection method, model generation method and terminal equipment |
| CN114022793A (en) * | 2021-10-28 | 2022-02-08 | 天津大学 | Optical remote sensing image change detection method based on twin network |
Non-Patent Citations (4)
| Title |
|---|
| 向阳等: "基于改进UNet孪生网络的遥感影像矿区变化检测", 《煤炭学报》 * |
| 吴纹辉等: "基于孪生网络和典型语义分割模型的遥感影像变化检测方法框架研究", 《国土资源信息化》 * |
| 张辉等: "基于多注意力图的孪生网络视觉目标跟踪", 《信号处理》 * |
| 薛白等: "基于孪生注意力网络的高分辨率遥感影像变化检测", 《自然资源遥感》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113077424A (en) * | 2021-03-23 | 2021-07-06 | 广东电网有限责任公司广州供电局 | Power transmission line channel environment change detection method and system based on deep learning |
| CN113506316A (en) * | 2021-05-27 | 2021-10-15 | 北京迈格威科技有限公司 | Method and device for segmenting video object and network model training method |
| CN113743298A (en) * | 2021-09-03 | 2021-12-03 | 云南电网有限责任公司电力科学研究院 | Power grid foreign matter detection method and device based on satellite image deep learning |
| CN113822867A (en) * | 2021-09-24 | 2021-12-21 | 东莞市诺丽电子科技有限公司 | Train appearance abnormity detection method and train appearance abnormity detection system |
| CN115330794A (en) * | 2022-10-13 | 2022-11-11 | 扬州中科半导体照明有限公司 | LED backlight foreign matter defect detection method based on computer vision |
| CN115330794B (en) * | 2022-10-13 | 2022-12-23 | 扬州中科半导体照明有限公司 | LED backlight foreign matter defect detection method based on computer vision |
| CN115620150A (en) * | 2022-12-05 | 2023-01-17 | 海豚乐智科技(成都)有限责任公司 | Multi-modal image ground building identification method and device based on twin transform |
| CN115620150B (en) * | 2022-12-05 | 2023-08-04 | 海豚乐智科技(成都)有限责任公司 | Multi-mode image ground building identification method and device based on twin transformers |
| CN118155082A (en) * | 2024-05-13 | 2024-06-07 | 山东锋士信息技术有限公司 | Double-branch change detection method based on hyperspectral space and spectral information |
| CN118410724A (en) * | 2024-07-02 | 2024-07-30 | 四川思极科技有限公司 | Transmission line foreign matter identification method, system, computer equipment and medium |
| CN118552800A (en) * | 2024-07-30 | 2024-08-27 | 广州英码信息科技有限公司 | Rail foreign matter semi-supervised anomaly detection method and system deployed at edge end |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112215085A (en) | Power transmission corridor foreign matter detection method and system based on twin network | |
| Chen et al. | Vehicle detection in high-resolution aerial images based on fast sparse representation classification and multiorder feature | |
| Chaudhuri et al. | Automatic building detection from high-resolution satellite images based on morphology and internal gray variance | |
| Lu et al. | Object-oriented change detection for landslide rapid mapping | |
| CN106683119B (en) | Moving vehicle detection method based on aerial video image | |
| CN110060273B (en) | Remote sensing image landslide mapping method based on deep neural network | |
| CN109829423B (en) | Infrared imaging detection method for frozen lake | |
| CN109063754A (en) | A kind of remote sensing image multiple features combining classification method based on OpenStreetMap | |
| CN109635726B (en) | Landslide identification method based on combination of symmetric deep network and multi-scale pooling | |
| CN112818905B (en) | Finite pixel vehicle target detection method based on attention and spatio-temporal information | |
| CN111832504A (en) | Space information intelligent integrated generation method for satellite in-orbit application | |
| CN110070545B (en) | Method for automatically extracting urban built-up area by urban texture feature density | |
| CN111291615A (en) | Multi-temporal remote sensing image change monitoring method | |
| CN108647658A (en) | A kind of infrared imaging detection method of high-altitude cirrus | |
| CN113838064A (en) | Cloud removing method using multi-temporal remote sensing data based on branch GAN | |
| CN110717531A (en) | Method for detecting classified change type based on uncertainty analysis and Bayesian fusion | |
| Zhang et al. | Multiple Saliency Features Based Automatic Road Extraction from High‐Resolution Multispectral Satellite Images | |
| CN118032789B (en) | Watch shell surface detection system and method based on machine vision | |
| Venkatakrishnamoorthy et al. | Cloud enhancement of NOAA multispectral images by using independent component analysis and principal component analysis for sustainable systems | |
| CN109829511B (en) | Texture classification-based method for detecting cloud layer area in downward-looking infrared image | |
| CN109785318B (en) | Remote sensing image change detection method based on facial line primitive association constraint | |
| CN112036246B (en) | Construction method of remote sensing image classification model, remote sensing image classification method and system | |
| CN115456886A (en) | Aviation remote sensing image shadow removing method based on deep learning and illumination model | |
| CN110610152B (en) | Multispectral cloud detection method based on discriminative feature learning unsupervised network | |
| CN114862883A (en) | Target edge extraction method, image segmentation method and system |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210112 |