CN115100540A - Method for automatically extracting high-resolution remote sensing image road - Google Patents
Method for automatically extracting high-resolution remote sensing image road Download PDFInfo
- Publication number
- CN115100540A CN115100540A CN202210778771.8A CN202210778771A CN115100540A CN 115100540 A CN115100540 A CN 115100540A CN 202210778771 A CN202210778771 A CN 202210778771A CN 115100540 A CN115100540 A CN 115100540A
- Authority
- CN
- China
- Prior art keywords
- road
- remote sensing
- agents
- extraction
- 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
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000000605 extraction Methods 0.000 claims abstract description 100
- 238000012549 training Methods 0.000 claims abstract description 15
- 238000013135 deep learning Methods 0.000 claims abstract description 13
- 238000012545 processing Methods 0.000 claims abstract description 6
- 238000004891 communication Methods 0.000 claims description 18
- 230000004927 fusion Effects 0.000 claims description 11
- 230000009471 action Effects 0.000 claims description 9
- 238000010586 diagram Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 239000011800 void material Substances 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 230000008447 perception Effects 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 10
- 238000003062 neural network model Methods 0.000 abstract description 4
- 230000008569 process Effects 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 4
- 230000011218 segmentation Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000001228 spectrum Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000002123 temporal effect Effects 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- 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
- G06V10/52—Scale-space analysis, e.g. wavelet analysis
-
- 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
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Remote Sensing (AREA)
- Astronomy & Astrophysics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of video image processing, in particular to a method for automatically extracting a high-resolution remote sensing image road, which comprises the steps of adding a multi-Agent system, creating and generating agents in different areas on an input high-resolution remote sensing image in the training process of a deep convolution neural network model, randomly generating a large number of extracting agents by the generating agents, enabling different extracting agents to be responsible for traversing different areas, adding a road map iteration generation algorithm model into the life cycle of each extracting Agent, and completing the road extraction task of the whole remote sensing image through information sharing and cooperation among different extracting agents, extracting agents and generating agents in the same area and among generating agents in different areas. The method realizes the high-efficiency training of the model through a multi-Agent technology, and can realize the high-efficiency and high-accuracy high-resolution remote sensing image road extraction by utilizing the strong characteristic learning capability of deep learning.
Description
Technical Field
The invention relates to the technical field of video image processing, in particular to a method for automatically extracting a high-resolution remote sensing image road.
Background
The remote sensing technology is a large-range earth observation technology with ultra-long distance perception, and can provide important data support for the development of economic society and the implementation of national major strategies by acquiring remote sensing images of interested areas and extracting and analyzing surface feature information. With the continuous development of remote sensing technology in recent years, remote sensing image data have characteristics of high spatial resolution, high spectral resolution and high temporal resolution, and although the high-resolution remote sensing images bring more abundant ground feature information, the difficulty of obtaining accurate required information from the high-resolution remote sensing images is increased. Therefore, how to accurately extract the required ground feature information from the high-resolution remote sensing image with high efficiency has become the current focus of research.
The road is a typical surface feature target with complex topological information in the remote sensing image, plays an important role in a plurality of scenes such as emergency response, traffic navigation, urban planning and the like, and the road information can also provide prior knowledge for identifying surface feature targets such as buildings, vegetation, rivers and the like in the task of understanding the remote sensing image scene. However, extracting road distribution from remote-sensed images is a challenging task due to the diversity and complexity of the road background environment in remote-sensed images. The traditional road extraction method of designing the feature extractor by means of manual experience can achieve a good extraction effect on a remote sensing image with a relatively simple scene, but is difficult to deal with interference caused by same-spectrum foreign matters and same-object different-spectrum phenomena, and is difficult to generate a high-quality road extraction result to be applied to an actual scene. Especially, with the continuous improvement of the resolution of the remote sensing image, the effect of the traditional road extraction method for non-deep learning is strong due to the interference factors such as illumination influence, tree and building shielding and the like.
With the cross combination of artificial intelligence and the remote sensing field, the deep learning method is applied to the identification and extraction of the ground object target of the remote sensing image and achieves better results. The convolutional neural network can automatically learn the rule of mapping the original input to the designated label, the learning capability gradually replaces the traditional mode of designing the characteristics by relying on artificial experience, the convolutional neural network becomes a mainstream algorithm model in the image processing fields of image classification, target detection, semantic segmentation and the like, and a new thought is provided for the high-resolution remote sensing image road extraction. The current deep learning algorithm for extracting roads surrounding high-resolution remote sensing images can be mainly divided into a semantic segmentation method based on pixels and a topological tracking method based on graph iteration. The former is characterized by predicting classes pixel by pixel, the principle is easy to operate but the efficiency is low, and the latter is characterized by searching and constructing a road network in an iterative manner by taking the topological shape of a road as a main body, so that the network model design is relatively complex but the road extraction effect is better. If the two methods can be further combined to realize advantage complementation, a new breakthrough can be made in the field of high-resolution remote sensing image road extraction, and the landing of the remote sensing image road extraction in practical application is promoted.
Disclosure of Invention
In order to solve the problems in the prior art, the invention aims to provide a method for automatically extracting a high-resolution remote sensing image road, and aims to solve the problems of low training efficiency and poor road extraction result of the current high-resolution remote sensing image road extraction model.
A method for automatically extracting a high-resolution remote sensing image road comprises the following steps:
acquiring a data set consisting of high-resolution remote sensing images accurately labeled by road category labels;
preprocessing the high-resolution remote sensing image in the data set;
training the road extraction model:
firstly, initializing a preprocessed high-resolution remote sensing image in different regions to generate a predetermined number of generating agents, allowing adjacent regions to coincide, and randomly initializing the generating agents in the regions in charge of the generating agents to generate a plurality of extracting agents;
each extraction Agent executes a road map iteration generation algorithm model based on deep learning to extract roads; in the life cycle of the whole roadmap iteration generation algorithm model, information sharing is kept among all extraction agents in the same region, between generation agents and extraction agents, and among generation agents in different regions, and the whole roadmap road extraction of the high-resolution remote sensing image is completed cooperatively;
after each extraction Agent is terminated, the stored road extraction results are required to be transmitted into a shared result library, and after all the extraction agents are terminated, all the road extraction results in the shared result library are combined to obtain a complete road extraction result graph;
carrying out loss calculation on the road extraction result graph and the corresponding real label graph, carrying out back propagation, and updating parameters of an algorithm model generated by iteration of the road graph; repeatedly training for multiple times to obtain a final road extraction model; and realizing automatic extraction of the road based on the final road extraction model.
On the basis of the current high-resolution remote sensing image road iteration generation algorithm with a single starting point of a whole image, a multi-Agent technology is introduced; in the multi-Agent system, the generating agents are responsible for generating a large number of extracting agents in different areas on an image, each extracting Agent is used as a starting point of a road map iteration generating algorithm model, and road extraction is carried out in the area to which the extracting Agent belongs in the life cycle of each extracting Agent; meanwhile, information sharing is kept among extraction agents within a certain range, and the road extraction work of the high-resolution remote sensing image is efficiently completed.
Preferably, the extracting Agent comprises a first sensing module, a first decision module, a first recording module, an action module and a first communication module;
the first sensing module cuts a remote sensing image with a preset size by taking the position of the extraction Agent as a center, and transmits the cut remote sensing image to the first decision module;
the first decision module comprises a road map iteration generation algorithm model based on deep learning, and the road map iteration generation algorithm model outputs the position of the next pixel point belonging to the road category according to the cut remote sensing image and transmits the position to the action module;
the action module is used for guiding the extraction Agent to move to the next predicted pixel point position belonging to the road category;
the first recording module adds the predicted position of the next pixel point belonging to the road category to the separately maintained road map and adds the corresponding edge;
the first communication module is used for guiding information sharing between the first communication module and other extraction agents in the same area and the generation agents of the area, and storing the road extraction result into a sharing result library when the information sharing is stopped.
Preferably, the preprocessing of the data includes performing data enhancement operation on the high-resolution remote sensing image and the road real label map corresponding to the high-resolution remote sensing image according to a predetermined probability, and uniformly adjusting the size of the high-resolution remote sensing image subjected to the data enhancement operation to 512 × 512.
Preferably, the data enhancement operation includes saturation change, horizontal turning, vertical turning and the like of the high-resolution remote sensing image.
Preferably, each generating Agent is responsible for the creation and monitoring management of extracting agents in the region;
the generating Agent comprises a second sensing module, a second communication module, a second recording module and a second decision module;
the second sensing module receives the image information of the region, and the image information in the region is used as prior knowledge for creating an extraction Agent;
the second communication module is responsible for communication with the generation agents of the adjacent area and all the extraction agents in the area, and information sharing is achieved;
the second recording module is responsible for recording the active states of all extraction agents in the region, and when all the extraction agents are terminated, the generated agents enter a suspended state to wait for a next creation command;
and the second decision module is responsible for random generation and collision avoidance of extraction agents in the region.
Preferably, the specific steps of obtaining the position of the next pixel point belonging to the road category by the road map iterative generation algorithm model are as follows:
the method comprises the steps that an Agent is extracted, remote sensing images with the size of H multiplied by W multiplied by 3 are intercepted by taking the position of the Agent as a center and sent into a coding network, a characteristic diagram of H 'multiplied by W' multiplied by C is obtained after down-sampling processing of the coding network, the characteristic diagram is respectively sent into a multi-scale cavity convolution fusion module and a space attention module which are in parallel, first road characteristic information and second road characteristic information are obtained, and the first road characteristic information and the second road characteristic information are merged and then sent into a decoding network; and obtaining the position of the next pixel point belonging to the road category through a decoding network.
Preferably, the training of the road map iterative generation algorithm model constructs a loss function by combining binary cross entropy loss and Dice loss, as follows:
Loss()＝BCE_Loss()+Soft_Dice_Loss()。
preferably, the multi-scale void convolution fusion module acquires the road characteristic information in a manner of cascading a plurality of convolution modules with different void convolution rates.
Preferably, the spatial attention module acquires global attention information of the road in the input image through a single cross spatial attention module; the calculation in the cross-space attention module is as follows:
H′＝∑A i Φ i +H；
in the formula: a. the i Representing the extracted attention feature, Φ i And (5) representing the feature information extracted from the input feature graph H, and finally outputting a feature graph H' after the cross attention feature is strengthened.
Because the high-resolution remote sensing image contains more pixel points and the pixel points belonging to the road are fewer, the neural network model of the road map iterative algorithm model is designed by adopting a coding-decoding structure, and adding a multi-scale cavity convolution fusion module and a cross space attention module which are parallel to each other, so that the extraction capability of the network model for the road information is improved.
The beneficial effects of the invention include:
on the basis of the current high-resolution remote sensing image road iteration generation algorithm with a single starting point of a whole image, a multi-Agent technology is introduced; in the multi-Agent system, the generating agents are responsible for generating a large number of extracting agents in different areas on an image, each extracting Agent is used as a starting point of a road map iteration generating algorithm model, and road extraction is carried out in the area to which the extracting Agent belongs in the life cycle of each extracting Agent; meanwhile, information sharing is kept among extraction agents within a certain range, and the road extraction work of the high-resolution remote sensing image is efficiently completed.
Because the high-resolution remote sensing image contains more pixel points and the pixel points belonging to the road are fewer, the neural network model of the road map iterative algorithm model is designed by adopting a coding-decoding structure and adding a multi-scale cavity convolution fusion module and a cross space attention module which are parallel to each other, so that the extraction capability of the network model on the road information is improved.
The method for extracting the high-resolution remote sensing image road is based on a topological tracking method based on graph iteration, a global road semantic segmentation prediction result is added as supplementary information during each iteration exploration, and the accuracy of road extraction is improved by combining the global road semantic segmentation prediction result and the supplementary information. In addition, the existing high-resolution remote sensing image road extraction method based on graph iteration starts iterative exploration from a single starting point, and the time consumption for traversing the whole graph is long, so that the model training efficiency is low. The method creatively combines a multi-Agent technology and a road extraction method based on roadmap iteration generation, not only realizes the high-efficiency training of the model through the multi-Agent technology, but also can realize the high-efficiency and high-accuracy road extraction of the high-resolution remote sensing image by utilizing the strong characteristic learning capability of deep learning.
Drawings
FIG. 1 is a basic flow chart of a high-resolution remote sensing image road automatic extraction method based on a multi-Agent technology and deep learning in the invention.
FIG. 2 is a design structure diagram of a multi-Agent system constructed by the present invention.
FIG. 3 is a model structure diagram of a deep learning-based road map iterative algorithm constructed by the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all the embodiments. The components of embodiments of the present application, generally described and illustrated in the figures herein, may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present application, presented in the accompanying drawings, is not intended to limit the scope of the claimed application, but is merely representative of selected embodiments of the application. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present application without making any creative effort, shall fall within the protection scope of the present application.
The following describes embodiments of the present invention in further detail with reference to the accompanying fig. 1:
a method for automatically extracting a high-resolution remote sensing image road comprises the following steps:
acquiring a data set consisting of high-resolution remote sensing images accurately labeled by road category labels;
preprocessing the high-resolution remote sensing image in the data set; the data preprocessing comprises the steps of carrying out data enhancement operation on the high-resolution remote sensing image and the road real label graph corresponding to the high-resolution remote sensing image according to a preset probability, and uniformly adjusting the size of the high-resolution remote sensing image subjected to the data enhancement operation to 512 x 512; and the data enhancement operation comprises saturation change, horizontal turning, vertical turning and the like of the high-resolution remote sensing image.
Training the road extraction model:
firstly, initializing a preprocessed high-resolution remote sensing image in different regions to generate a predetermined number of generating agents, allowing adjacent regions to coincide, and randomly initializing the generating agents in the regions in charge of the generating agents to generate a plurality of extracting agents;
each extraction Agent executes a road map iteration generation algorithm model based on deep learning to extract roads; in the life cycle of the whole roadmap iteration generation algorithm model, information sharing is kept among all extraction agents in the same region, between generation agents and extraction agents, and among generation agents in different regions, and the whole roadmap road extraction of the high-resolution remote sensing image is completed cooperatively; after each extraction Agent is terminated, the stored road extraction results are required to be transmitted into a shared result library, and after all the extraction agents are terminated, all the road extraction results in the shared result library are combined to obtain a complete road extraction result graph;
performing loss calculation on the road extraction result graph and the corresponding real label graph, performing back propagation, and updating parameters of an algorithm model generated by iteration of the road graph; repeatedly training for multiple times to obtain a final road extraction model; and realizing automatic extraction of the road based on the final road extraction model.
On the basis of the current high-resolution remote sensing image road iteration generation algorithm with a whole picture and a single starting point, a multi-Agent technology is introduced; in the multi-Agent system, the generating agents are responsible for generating a large number of extracting agents in different areas on an image, each extracting Agent is used as a starting point of a road map iteration generating algorithm model, and road extraction is carried out in the area to which the extracting Agent belongs in the life cycle of each extracting Agent; meanwhile, information sharing is kept among extraction agents within a certain range, and the road extraction work of the high-resolution remote sensing image is efficiently completed.
Referring to fig. 2, the extracting Agent includes a first sensing module, a first decision module, a first recording module, an action module, and a first communication module;
the first sensing module cuts a remote sensing image with a preset size by taking the position of the extraction Agent as a center, and transmits the cut remote sensing image to the first decision module;
the first decision module comprises a road map iteration generation algorithm model based on deep learning, the road map iteration generation algorithm model outputs the position of a next pixel point belonging to the road category according to the cut remote sensing image and transmits the position to the action module;
the action module is used for guiding the extraction Agent to move to the next predicted pixel point position belonging to the road category;
the first recording module adds the predicted position of the next pixel point belonging to the road category to the separately maintained road map and adds the corresponding edge; considering that the initial position of the extraction Agent during generation is not necessarily on the road, the initial position is not added into the road map;
the first communication module is used for guiding information sharing between the first communication module and other extraction agents in the same area and the generation agents of the area, and storing the road extraction result into a sharing result library when the information sharing is stopped.
Referring to fig. 2, each generating Agent is responsible for creation and monitoring management of extracting agents in the region;
the generating Agent comprises a second sensing module, a second communication module, a second recording module and a second decision module;
the second sensing module receives the image information of the region, and the image information in the region is used as prior knowledge for creating an extraction Agent;
the second communication module is responsible for communication with the generation agents of the adjacent area and all the extraction agents in the area, and information sharing is achieved;
the second recording module is responsible for recording the active states of all the extraction agents in the region, and when all the extraction agents are terminated, the Agent entering and hanging state is generated to wait for the next creation command;
the second decision module is responsible for random generation and collision avoidance of extraction agents in the region;
the process is repeated continuously in the life cycle of each generation Agent until all extraction agents in the area are terminated.
Referring to fig. 3, the specific steps of obtaining the position of the next pixel point belonging to the road category by the road map iterative generation algorithm model are as follows:
the method comprises the steps that an Agent is extracted, remote sensing images with the size of H multiplied by W multiplied by 3 are intercepted by taking the position of the Agent as a center and sent into a coding network, a characteristic diagram of H 'multiplied by W' multiplied by C is obtained after down-sampling processing of the coding network, the characteristic diagram is respectively sent into a multi-scale cavity convolution fusion module and a space attention module which are in parallel, first road characteristic information and second road characteristic information are obtained, and the first road characteristic information and the second road characteristic information are merged and then sent into a decoding network; and obtaining the position of the next pixel point belonging to the road category through a decoding network.
Considering the sample imbalance problem between the road and the background in the training sample, the training of the road map iterative generation algorithm model constructs a loss function by combining binary cross entropy loss and Dice loss, as follows:
Loss()＝BCE_Loss()+Soft_Dice_Loss()。
the multi-scale void convolution fusion module acquires road characteristic information in a mode of cascading a plurality of convolution modules with different void convolution rates.
The space attention module acquires global attention information of a road in the input image through the two cross space attention modules; the calculation in the cross-space attention module is as follows:
H′＝∑A i Φ i +H；
in the formula: a. the i Representing the extracted attention feature, Φ i And (5) representing the feature information extracted from the input feature graph H, and finally outputting a feature graph H' after the cross attention feature is strengthened.
By combining the multi-scale cavity convolution fusion module and the space attention module, the defect that only local information is emphasized in the existing method is overcome, and the extraction effect of the road characteristics is improved by combining the multi-scale cavity convolution fusion module and the space attention module.
Because the high-resolution remote sensing image contains more pixel points and the pixel points belonging to the road are fewer, the neural network model of the road map iterative algorithm model is designed by adopting a coding-decoding structure, and adding a multi-scale cavity convolution fusion module and a cross space attention module which are parallel to each other, so that the extraction capability of the network model for the road information is improved.
The above-mentioned embodiments only express the specific embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for those skilled in the art, without departing from the technical idea of the present application, several changes and modifications can be made, which are all within the protection scope of the present application.
Claims (9)
1. A method for automatically extracting a high-resolution remote sensing image road is characterized by comprising the following steps:
acquiring a data set consisting of high-resolution remote sensing images accurately labeled by road category labels;
preprocessing the high-resolution remote sensing image in the data set;
training the road extraction model:
firstly, initializing a preprocessed high-resolution remote sensing image in different areas to generate a preset number of generating agents, allowing adjacent areas to coincide, and randomly initializing the generating agents in the areas responsible for the generating agents to generate a plurality of extracting agents;
each extraction Agent executes a road map iteration generation algorithm model based on deep learning to extract roads; in the life cycle of the whole road map iteration generation algorithm model, information sharing is kept among all extraction agents in the same region, between generation agents and extraction agents, and among generation agents in different regions, and the whole-map road extraction of the high-resolution remote sensing image is completed cooperatively;
after each extraction Agent is terminated, the stored road extraction results are required to be transmitted into a shared result library, and after all the extraction agents are terminated, all the road extraction results in the shared result library are combined to obtain a complete road extraction result graph;
performing loss calculation on the road extraction result graph and the corresponding real label graph, performing back propagation, and updating parameters of an algorithm model generated by iteration of the road graph; repeatedly training for multiple times to obtain a final road extraction model; and realizing automatic extraction of the road based on the final road extraction model.
2. The method for automatically extracting the high-resolution remote sensing image road according to claim 1, wherein the extraction Agent comprises a first perception module, a first decision module, a first recording module, an action module and a first communication module;
the first sensing module cuts a remote sensing image with a preset size by taking the position of the extraction Agent as a center, and transmits the cut remote sensing image to the first decision module;
the first decision module comprises a road map iteration generation algorithm model based on deep learning, the road map iteration generation algorithm model outputs the position of a next pixel point belonging to the road category according to the cut remote sensing image and transmits the position to the action module;
the action module is used for guiding the extraction Agent to move to the next predicted pixel point position belonging to the road category;
the first recording module adds the predicted position of the next pixel point belonging to the road category to the separately maintained road map and adds the corresponding edge;
the first communication module is used for guiding information sharing between the first communication module and other extraction agents in the same area and the generation agents of the area, and storing the road extraction result into a sharing result library when the information sharing is stopped.
3. The method according to claim 1, wherein the preprocessing of the data comprises performing data enhancement operation on the high-resolution remote sensing image and a road real label map corresponding to the high-resolution remote sensing image according to a predetermined probability, and uniformly adjusting the size of the high-resolution remote sensing image subjected to the data enhancement operation to 512 x 512.
4. The method according to claim 3, wherein the data enhancement operation comprises saturation change, horizontal inversion and vertical inversion of the high-resolution remote sensing image.
5. The method for automatically extracting the high-resolution remote sensing image road according to claim 1, wherein each generating Agent is responsible for creation and monitoring management of extracting agents in the region;
the generating Agent comprises a second sensing module, a second communication module, a second recording module and a second decision module;
the second sensing module receives the image information of the region, and the image information in the region is used as priori knowledge for creating and extracting the Agent;
the second communication module is responsible for communication with the generation agents of the adjacent area and all the extraction agents in the area, and information sharing is achieved;
the second recording module is responsible for recording the active states of all the extracting agents in the region, and when all the extracting agents are terminated, the generating agents enter a suspended state to wait for the next creation command;
and the second decision module is responsible for random generation and collision avoidance of extraction agents in the region.
6. The method for automatically extracting the high-resolution remote sensing image road according to claim 1, wherein the specific steps of obtaining the next pixel point position belonging to the road category by the road map iteration generation algorithm model are as follows:
the method comprises the steps that an Agent is extracted, remote sensing images with the size of H multiplied by W multiplied by 3 are intercepted by taking the position of the Agent as a center and sent into a coding network, a characteristic diagram of H 'multiplied by W' multiplied by C is obtained after down-sampling processing of the coding network, the characteristic diagram is respectively sent into a multi-scale cavity convolution fusion module and a space attention module which are in parallel, first road characteristic information and second road characteristic information are obtained, and the first road characteristic information and the second road characteristic information are merged and then sent into a decoding network; and obtaining the position of the next pixel point belonging to the road category through a decoding network.
7. The method for automatically extracting the high-resolution remote sensing image road according to claim 1, wherein the training of the road map iterative generation algorithm model constructs a loss function by combining binary cross entropy loss and Dice loss as follows:
Loss()＝BCE_Loss()+Soft_Dice_Loss()。
8. the method for automatically extracting the high-resolution remote sensing image road according to claim 6, wherein the multi-scale void convolution fusion module acquires road characteristic information in a mode that a plurality of convolution modules with different void convolution rates are cascaded.
9. The method for automatically extracting the high-resolution remote sensing image road according to claim 6, wherein the spatial attention module acquires global attention information of the road in the input image through two cross spatial attention modules; the calculation in the cross space attention module is as follows:
H′＝∑A i Φ i +H；
in the formula: a. the i Representing the extracted attention feature, Φ i And (5) representing the feature information extracted from the input feature graph H, and finally outputting a feature graph H' after the cross attention feature is strengthened.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210778771.8A CN115100540B (en) | 2022-06-30 | 2022-06-30 | Automatic road extraction method for high-resolution remote sensing image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210778771.8A CN115100540B (en) | 2022-06-30 | 2022-06-30 | Automatic road extraction method for high-resolution remote sensing image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115100540A true CN115100540A (en) | 2022-09-23 |
| CN115100540B CN115100540B (en) | 2024-05-07 |
Family
ID=83294221
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210778771.8A Active CN115100540B (en) | 2022-06-30 | 2022-06-30 | Automatic road extraction method for high-resolution remote sensing image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115100540B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090043486A1 (en) * | 2007-07-27 | 2009-02-12 | Chaowei Yang | Near Real-time Traffic Routing |
| CN105007191A (en) * | 2015-07-31 | 2015-10-28 | 上海红神信息技术有限公司 | Perception, decision and execution separation method for pseudo-route switching system |
| CN107480726A (en) * | 2017-08-25 | 2017-12-15 | 电子科技大学 | A kind of Scene Semantics dividing method based on full convolution and shot and long term mnemon |
| CN107729922A (en) * | 2017-09-20 | 2018-02-23 | 千寻位置网络有限公司 | Remote sensing images method for extracting roads based on deep learning super-resolution technique |
| WO2018132607A2 (en) * | 2017-01-12 | 2018-07-19 | Mobileye Vision Technologies Ltd. | Navigation based on vehicle activity |
| CN110659342A (en) * | 2019-08-21 | 2020-01-07 | 北京吉威数源信息技术有限公司 | Rule base automatic drawing comprehensive processing method and system based on information extraction features |
| WO2021184891A1 (en) * | 2020-03-20 | 2021-09-23 | 中国科学院深圳先进技术研究院 | Remotely-sensed image-based terrain classification method, and system |
| CN113591608A (en) * | 2021-07-12 | 2021-11-02 | 浙江大学 | High-resolution remote sensing image impervious surface extraction method based on deep learning |
| CN114581748A (en) * | 2022-05-06 | 2022-06-03 | 南京大学 | Multi-agent perception fusion system based on machine learning and implementation method thereof |
-
2022
- 2022-06-30 CN CN202210778771.8A patent/CN115100540B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090043486A1 (en) * | 2007-07-27 | 2009-02-12 | Chaowei Yang | Near Real-time Traffic Routing |
| CN105007191A (en) * | 2015-07-31 | 2015-10-28 | 上海红神信息技术有限公司 | Perception, decision and execution separation method for pseudo-route switching system |
| WO2018132607A2 (en) * | 2017-01-12 | 2018-07-19 | Mobileye Vision Technologies Ltd. | Navigation based on vehicle activity |
| CN107480726A (en) * | 2017-08-25 | 2017-12-15 | 电子科技大学 | A kind of Scene Semantics dividing method based on full convolution and shot and long term mnemon |
| CN107729922A (en) * | 2017-09-20 | 2018-02-23 | 千寻位置网络有限公司 | Remote sensing images method for extracting roads based on deep learning super-resolution technique |
| CN110659342A (en) * | 2019-08-21 | 2020-01-07 | 北京吉威数源信息技术有限公司 | Rule base automatic drawing comprehensive processing method and system based on information extraction features |
| WO2021184891A1 (en) * | 2020-03-20 | 2021-09-23 | 中国科学院深圳先进技术研究院 | Remotely-sensed image-based terrain classification method, and system |
| CN113591608A (en) * | 2021-07-12 | 2021-11-02 | 浙江大学 | High-resolution remote sensing image impervious surface extraction method based on deep learning |
| CN114581748A (en) * | 2022-05-06 | 2022-06-03 | 南京大学 | Multi-agent perception fusion system based on machine learning and implementation method thereof |
Non-Patent Citations (4)
| Title |
|---|
| RENBAO LIAN等: "Road Extraction Methods in High-Resolution Remote Sensing Images: A Comprehensive Review", 《IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING》, 11 September 2020 (2020-09-11) * |
| 戴激光;杜阳;方鑫鑫;王杨;苗志鹏;: "多特征约束的高分辨率光学遥感影像道路提取", 遥感学报, no. 05, 25 September 2018 (2018-09-25) * |
| 王琪: "基于深度学习的高分辨率遥感影像道路提取方法研究", 《万方数据》, 2 December 2022 (2022-12-02) * |
| 赵贝;钟燕飞;张良培;: "高空间分辨率遥感影像的多智能体分割方法研究", 测绘学报, no. 01, 15 February 2013 (2013-02-15) * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115100540B (en) | 2024-05-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113298818B (en) | Remote sensing image building segmentation method based on attention mechanism and multi-scale features | |
| CN111046768B (en) | Deep learning method for simultaneously extracting road pavement and center line of remote sensing image | |
| CN111814683A (en) | Robust visual SLAM method based on semantic prior and deep learning features | |
| CN113780296A (en) | Remote sensing image semantic segmentation method and system based on multi-scale information fusion | |
| CN110879960B (en) | Method and computing device for generating image data set for convolutional neural network learning | |
| CN113344932B (en) | Semi-supervised single-target video segmentation method | |
| CN115879535A (en) | Training method, device, equipment and medium for automatic driving perception model | |
| CN113076871A (en) | Fish shoal automatic detection method based on target shielding compensation | |
| CN114519932B (en) | Regional traffic condition integrated prediction method based on space-time relation extraction | |
| CN115860102B (en) | Pre-training method, device, equipment and medium for automatic driving perception model | |
| CN114943876A (en) | Cloud and cloud shadow detection method and device for multi-level semantic fusion and storage medium | |
| CN110378284B (en) | Road front view generation method and device, electronic equipment and storage medium | |
| CN115512169A (en) | Weak supervision semantic segmentation method and device based on gradient and region affinity optimization | |
| Buyukdemircioglu et al. | Deep learning for 3D building reconstruction: A review | |
| CN114511682A (en) | Three-dimensional scene reconstruction method and device based on laser radar and electronic equipment | |
| CN104463962A (en) | Three-dimensional scene reconstruction method based on GPS information video | |
| Vödisch et al. | CoDEPS: Online continual learning for depth estimation and panoptic segmentation | |
| Nguyen | An efficient license plate detection approach using lightweight deep convolutional neural networks | |
| CN103473789A (en) | Human body video segmentation method fusing multi-cues | |
| Lu et al. | OctreeOcc: Efficient and Multi-Granularity Occupancy Prediction Using Octree Queries | |
| Xiang et al. | Nisb-map: Scalable mapping with neural implicit spatial block | |
| CN115100540B (en) | Automatic road extraction method for high-resolution remote sensing image | |
| CN116503602A (en) | Unstructured environment three-dimensional point cloud semantic segmentation method based on multi-level edge enhancement | |
| CN116402874A (en) | Spacecraft depth complementing method based on time sequence optical image and laser radar data | |
| CN113514053B (en) | Method and device for generating sample image pair and method for updating high-precision map |
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 |