CN114913428A - Remote sensing image target detection system based on deep learning - Google Patents
Remote sensing image target detection system based on deep learning Download PDFInfo
- Publication number
- CN114913428A CN114913428A CN202210446366.6A CN202210446366A CN114913428A CN 114913428 A CN114913428 A CN 114913428A CN 202210446366 A CN202210446366 A CN 202210446366A CN 114913428 A CN114913428 A CN 114913428A
- Authority
- CN
- China
- Prior art keywords
- detection
- remote sensing
- sensing image
- polar
- target
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 41
- 238000013135 deep learning Methods 0.000 title claims abstract description 9
- 230000004927 fusion Effects 0.000 claims abstract description 6
- 230000007246 mechanism Effects 0.000 claims abstract description 6
- 238000010586 diagram Methods 0.000 claims description 5
- 238000000034 method Methods 0.000 abstract description 10
- 230000008859 change Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000023004 detection of visible light Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000003703 image analysis method Methods 0.000 description 1
- 238000002347 injection Methods 0.000 description 1
- 239000007924 injection Substances 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000000007 visual 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
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
- G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- 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)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a remote sensing image target detection system based on deep learning, and relates to the technical field of image target detection; the detection method comprises the following steps: the module is integrated into a whole and is used for replacing a main network ResNet-101 in the original detection algorithm and inputting an image; a HRNet high resolution network; performing feature fusion on the FPN-like structure; a channel domain attention mechanism; a spatial domain attention mechanism; a feature map; predicting key points; predicting the polar diameter and the polar angle of the polar coordinate system; obtaining pole coordinates (x, y), and obtaining a polar diameter and a polar angle; finally outputting a target prediction frame; the method is suitable for detecting the target of the remote sensing image by integrating a plurality of detection difficulties of the remote sensing image, so as to realize quick detection, reduce the detection workload and shorten the time; accurate target detection can be realized, and the network detection precision is higher.
Description
Technical Field
The invention belongs to the technical field of image target detection, and particularly relates to a remote sensing image target detection system based on deep learning.
Background
The target detection task based on the remote sensing image is to identify and position specific objects in the image, and plays an important role in sea vessel control, environmental quality monitoring, ground planning and layout and the like. Therefore, how to improve the performance of target detection to obtain a detection result with more accurate positioning and more detailed classification becomes a key research content in the field. The primary condition of positioning and identification in the task of detecting the remote sensing image target is to extract the characteristics of the image. The traditional detection methods comprise a template matching method, a region prior-based method, an image analysis method and the like, all of which need manual feature design, have strong specificity of features and do not have cross-class universality.
Nowadays, the aerospace technology is rapidly developed in China, the remote sensing technology is continuously improved and matured, the remote sensing image is more convenient and efficient to obtain than before, and the method has the characteristics of large data volume, rich varieties, clear imaging, high timeliness and the like, and is very in line with the development trend of the current social big data era. Therefore, a solid foundation is laid for the development of the remote sensing image detection technology, and more valuable and abundant and diverse data information can be provided for the deep learning method. Currently, for remote sensing images, a deep learning target detection method can be migrated to the field of remote sensing detection. However, the remote sensing image has certain difference from the natural scene image, and the remote sensing image has unique characteristics such as high resolution, small and dense targets, complex background and the like due to different imaging modes. Therefore, the migration method needs to be improved in a targeted manner, so that the difficult problems in remote sensing image detection are solved, and overcoming the critical detection difficulties becomes a key research direction for many scholars in the field. The challenges of target detection of visible light remote sensing images are as follows: 1. with the improvement of resolution and the expansion of an imaging area, a large number of complex natural and artificial backgrounds exist in a visible light remote sensing image, and serious interference is generated on the detection of a target. 2. The direction change is large: the remote sensing image is shot from an aerial visual angle, a scene is a top view, targets are distributed in the scene at various angles, the adaptability of most of the existing algorithms to the angles is not high, and the robustness is not enough when the multi-direction problem is processed. In addition, when a multi-direction target is positioned by a classical horizontal frame positioning mode, a surrounding frame is not compact enough, and positioning is not fine enough. 3. The remote sensing image target has large scale change, in order to solve the problem, an anchor detector needs to be provided with anchors with various scales, and a large amount of low-quality positive sample anchors are introduced in the anchor matching process, so that the detection precision of the detector is influenced. 4. The design of the anchor in the detection method with the anchor depends on the experience of people seriously, the pertinence to the data set is very strong, the data set needs to be redesigned when being replaced, and the generalization is lacked.
Disclosure of Invention
To solve the problems in the background art; the invention aims to provide a remote sensing image target detection system based on deep learning.
The invention relates to a remote sensing image target detection system based on deep learning, which comprises the following detection methods: the module is fused into a whole and is used for replacing a backbone network ResNet-101, (1) and an input image in the original detection algorithm; (2) HRNet high resolution network; (3) carrying out feature fusion on the FPN-like structure; (4) attention mechanism of channel region; (5) spatial domain attention mechanism; (6) a characteristic diagram; (7): (7.1) predicting key points; (7.2) predicting the polar diameter and the polar angle of the polar coordinate system; (8): (8.1), obtaining pole coordinates (x, y), (8.2), and obtaining a polar diameter and a polar angle; (9) and finally outputting the target prediction frame.
Compared with the prior art, the invention has the beneficial effects that:
the method is suitable for detecting the target of the remote sensing image by integrating a plurality of detection difficulties of the remote sensing image, so that the rapid detection is realized, the detection workload is reduced, and the time is shortened.
Secondly, accurate target detection can be realized, and the network detection precision is higher.
Drawings
For ease of illustration, the invention is described in detail by the following detailed description and the accompanying drawings.
FIG. 1 is a schematic diagram of the original P-RSDet framework;
FIG. 2 is a flow chart of an improved schematic network structure of the present invention;
FIG. 3 is a schematic diagram of the overall network architecture of the present invention;
FIG. 4 is a schematic diagram of the structures of BottleNeck and BasicBlock in the present invention.
Detailed Description
In order that the objects, aspects and advantages of the invention will become more apparent, the invention will be described by way of example only, and in connection with the accompanying drawings. It is to be understood that this description is made only by way of example and not as a limitation on the scope of the invention. The structure, proportion, size and the like shown in the drawings are only used for matching with the content disclosed in the specification, so that the person skilled in the art can understand and read the description, and the description is not used for limiting the limit condition of the implementation of the invention, so the method has no technical essence, and any structural modification, proportion relation change or size adjustment still falls within the range covered by the technical content disclosed by the invention without affecting the effect and the achievable purpose of the invention. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present invention.
It should be noted that, in order to avoid obscuring the present invention with unnecessary details, only the structures and/or processing steps closely related to the scheme according to the present invention are shown in the drawings, and other details not so relevant to the present invention are omitted.
The specific implementation mode adopts the following technical scheme: as shown in fig. 1, a scheme for improving the detection accuracy is provided for the original frame P-RSDet keypoint detection method. And the module is fused into a whole and is used for replacing a backbone network ResNet-101 in the original detection algorithm (P-RSDet). The network structure before and after the improvement is expected: as shown in fig. 2 and 3;
in the feature extraction network shown in fig. 3, BottleNeck is used for the transformation of the internal feature map in block 1, see the left half of fig. 4, and BasicBlock is used for the transformation of the internal feature maps in blocks 2, 3, and 4, see the right half of fig. 4. The up-sampling uses bilinear interpolation followed by 1X1 convolution to adjust the number of channels, and the down-sampling uses 2 steps and 3X3 convolution kernel size. In an FPN (feature pyramid) -like structure feature fusion network, for a target with a small size, pyramid levels of a middle lower layer containing more image bottom layer information and detail information often have a better detection effect, the image layers have higher resolution, and the perception degree of the whole information of an image is also higher. It can be speculated that these pyramid levels of the lower partial layers will have more play space in the remote sensing image. Therefore, the generation mode of each level of the classical feature pyramid network is reserved, and branches are added on the basis of the classical feature pyramid network, namely, information of the pyramid at the upper layer is added in the pyramid levels at the middle and lower layers, namely, feature maps after the feature maps are sampled are fused to add more abstract semantic information. It is considered necessary to add high-level features on the basis of bottom-level information, and for complex scenes and huge target scale spans of remote sensing images, the fusion of high-level semantic information has good effect. Therefore, the low-level feature map not only contains the original bottom-layer image information and detail features, but also greatly enhances the perception capability of the target after the injection of the high-level semantic features. Through the fusion structure of the multilayer pyramid, the whole network has stronger adaptability to small-size targets and targets with large size change.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.
Claims (1)
1. A remote sensing image target detection system based on deep learning is characterized in that: the detection method comprises the following steps: the module is fused into a whole and is used for replacing a backbone network ResNet-101, (1) and an input image in the original detection algorithm; (2) HRNet high resolution network; (3) carrying out feature fusion on the FPN-like structure; (4) channel domain attention mechanism; (5) a spatial domain attention mechanism; (6) a characteristic diagram; (7): (7.1) predicting key points; (7.2) predicting the polar diameter and the polar angle of the polar coordinate system; (8): (8.1), obtaining pole coordinates (x, y), (8.2), and obtaining a polar diameter and a polar angle; (9) and finally outputting the target prediction frame.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210446366.6A CN114913428A (en) | 2022-04-26 | 2022-04-26 | Remote sensing image target detection system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210446366.6A CN114913428A (en) | 2022-04-26 | 2022-04-26 | Remote sensing image target detection system based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114913428A true CN114913428A (en) | 2022-08-16 |
Family
ID=82764252
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210446366.6A Pending CN114913428A (en) | 2022-04-26 | 2022-04-26 | Remote sensing image target detection system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114913428A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117115641A (en) * | 2023-07-20 | 2023-11-24 | 中国科学院空天信息创新研究院 | Building information extraction method and device, electronic equipment and storage medium |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111797697A (en) * | 2020-06-10 | 2020-10-20 | 河海大学 | Angle high-resolution remote sensing image target detection method based on improved CenterNet |
| CN111814771A (en) * | 2020-09-04 | 2020-10-23 | 支付宝(杭州)信息技术有限公司 | Image processing method and device |
| CN112070069A (en) * | 2020-11-10 | 2020-12-11 | 支付宝(杭州)信息技术有限公司 | Method and device for identifying remote sensing image |
| CN112837330A (en) * | 2021-03-02 | 2021-05-25 | 中国农业大学 | Leaf segmentation method based on multi-scale double attention mechanism and full convolution neural network |
| WO2021244079A1 (en) * | 2020-06-02 | 2021-12-09 | 苏州科技大学 | Method for detecting image target in smart home environment |
| CN114092820A (en) * | 2022-01-20 | 2022-02-25 | 城云科技（中国）有限公司 | Target detection method and moving target tracking method applying same |
| CN114220015A (en) * | 2021-12-21 | 2022-03-22 | 一拓通信集团股份有限公司 | Improved YOLOv 5-based satellite image small target detection method |
| CN114331949A (en) * | 2021-09-29 | 2022-04-12 | 腾讯科技（上海）有限公司 | Image data processing method, computer equipment and readable storage medium |
-
2022
- 2022-04-26 CN CN202210446366.6A patent/CN114913428A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021244079A1 (en) * | 2020-06-02 | 2021-12-09 | 苏州科技大学 | Method for detecting image target in smart home environment |
| CN111797697A (en) * | 2020-06-10 | 2020-10-20 | 河海大学 | Angle high-resolution remote sensing image target detection method based on improved CenterNet |
| CN111814771A (en) * | 2020-09-04 | 2020-10-23 | 支付宝(杭州)信息技术有限公司 | Image processing method and device |
| CN112070069A (en) * | 2020-11-10 | 2020-12-11 | 支付宝(杭州)信息技术有限公司 | Method and device for identifying remote sensing image |
| CN112837330A (en) * | 2021-03-02 | 2021-05-25 | 中国农业大学 | Leaf segmentation method based on multi-scale double attention mechanism and full convolution neural network |
| CN114331949A (en) * | 2021-09-29 | 2022-04-12 | 腾讯科技（上海）有限公司 | Image data processing method, computer equipment and readable storage medium |
| CN114220015A (en) * | 2021-12-21 | 2022-03-22 | 一拓通信集团股份有限公司 | Improved YOLOv 5-based satellite image small target detection method |
| CN114092820A (en) * | 2022-01-20 | 2022-02-25 | 城云科技（中国）有限公司 | Target detection method and moving target tracking method applying same |
Non-Patent Citations (3)
| Title |
|---|
| LIN ZHOU等: "Arbitrary-Oriented Object Detection in Remote Sensing Images Based on Polar Coordinates", 《IEEE ACCESS》, 28 December 2020 (2020-12-28), pages 1 - 12 * |
| XU HE等: "High-Resolution Polar Network for Object Detection in Remote Sensing Images", 《IEEE GEOSCIENCE AND REMOTE SENSING LETTERS》, 29 December 2021 (2021-12-29), pages 1 - 5 * |
| 李辉等: "面向输电线路的异常目标检测方法", 《计算机与现代化》, 31 December 2020 (2020-12-31), pages 1 - 7 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117115641A (en) * | 2023-07-20 | 2023-11-24 | 中国科学院空天信息创新研究院 | Building information extraction method and device, electronic equipment and storage medium |
| CN117115641B (en) * | 2023-07-20 | 2024-03-22 | 中国科学院空天信息创新研究院 | Building information extraction method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110084210B (en) | SAR image multi-scale ship detection method based on attention pyramid network | |
| CN110246181B (en) | Anchor point-based attitude estimation model training method, attitude estimation method and system | |
| CN111582104B (en) | Remote sensing image semantic segmentation method and device based on self-attention feature aggregation network | |
| CN111460984A (en) | Global lane line detection method based on key point and gradient balance loss | |
| US20080235184A1 (en) | Image Search Method, Image Search Apparatus, and Recording Medium Having Image Search Program Code Thereon | |
| CN111667030B (en) | Method, system and storage medium for realizing remote sensing image target detection based on deep neural network | |
| CN113435269A (en) | Improved water surface floating object detection and identification method and system based on YOLOv3 | |
| Chen et al. | ASF-Net: Adaptive screening feature network for building footprint extraction from remote-sensing images | |
| CN115223017B (en) | Multi-scale feature fusion bridge detection method based on depth separable convolution | |
| CN112818777B (en) | Remote sensing image target detection method based on dense connection and feature enhancement | |
| CN113255589A (en) | Target detection method and system based on multi-convolution fusion network | |
| Shi et al. | CloudU-Netv2: A cloud segmentation method for ground-based cloud images based on deep learning | |
| CN116343043B (en) | Remote sensing image change detection method with multi-scale feature fusion function | |
| CN114913428A (en) | Remote sensing image target detection system based on deep learning | |
| Wilson et al. | Image and object Geo-localization | |
| CN116719031B (en) | Ocean vortex detection method and system for synthetic aperture radar SAR image | |
| Zhang et al. | Multilevel feature context semantic fusion network for cloud and cloud shadow segmentation | |
| CN116503750A (en) | Large-range remote sensing image rural block type residential area extraction method and system integrating target detection and visual attention mechanisms | |
| Gong et al. | Crude Oil Leakage Detection Based on DA‐SR Framework | |
| Seo et al. | Sensor-rich video exploration on a map interface | |
| Zhang et al. | Feature enhanced centernet for object detection in remote sensing images | |
| CN114463203A (en) | Data enhancement method for small-scale target | |
| CN115082778A (en) | Multi-branch learning-based homestead identification method and system | |
| CN110196638B (en) | Mobile terminal augmented reality method and system based on target detection and space projection | |
| CN110377791A (en) | Fast image sort method |
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 |