CN114283322A - SAR image ship detection method - Google Patents
SAR image ship detection method Download PDFInfo
- Publication number
- CN114283322A CN114283322A CN202111606636.7A CN202111606636A CN114283322A CN 114283322 A CN114283322 A CN 114283322A CN 202111606636 A CN202111606636 A CN 202111606636A CN 114283322 A CN114283322 A CN 114283322A
- Authority
- CN
- China
- Prior art keywords
- ship
- image
- area
- detection
- sar 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.)
- Pending
Links
Images
Landscapes
- Image Processing (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The imaging width of the satellite-borne SAR is usually very wide, and the resolution is often not high enough, so that the imaging area of a sea ship on the SAR image is limited, and the ship target identification precision is influenced. In order to overcome the defects in the prior art, the invention provides the SAR image ship detection method, which not only can accurately obtain the position, the direction and the contour information of a ship, but also can accurately estimate the geometric characteristics, the shape characteristics, the edge characteristics and other characteristic information of the ship.
Description
Technical Field
The invention belongs to the field of radar detection, and particularly relates to an SAR image ship detection method.
Background
The imaging width of the satellite-borne SAR is usually very wide, and the resolution is often not high enough, so that the imaging area of a sea ship on the SAR image is limited, and the ship target identification precision is influenced. In order to solve the problem of SMALL-scale SAR ship target CLASSIFICATION, the 'JOINT CONVOLUONAL NEURAL NETWORK FOR SMALL-SCALE SHIP CLASSIFIFICATION IN SAR IMAGES' ('IGARSS') discloses a CONVOLUTIONAL NEURAL NETWORK which combines a generator NETWORK and a CLASSIFICATION NETWORK, wherein the generator NETWORK is used FOR learning a mapping function FOR converting a low-resolution ship image into a high-resolution ship image, and the CLASSIFICATION NETWORK classifies the generated high-resolution ship image as input. A joint loss optimization strategy is provided for training and optimizing the model, and experiments show that the ship picture generated by the generator effectively improves the classification precision of the model. However, the existing method has the disadvantages that the traditional convolutional neural network model is not specially designed for the ship detection frame, so that the detection frame is generally a forward rectangular frame in the result, only the approximate position of the ship can be estimated, the contour and the direction of the ship are difficult to accurately detect, and characteristic information such as the geometry, the shape, the edge, the scattering and the like of the ship cannot be extracted.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention provides the SAR image ship detection method, which not only can accurately obtain the position, the direction and the contour information of a ship, but also can accurately estimate the geometric characteristics, the shape characteristics, the edge characteristics and other characteristic information of the ship. The method comprises the following specific steps:
(1) image pre-processing
Image enhancement: the brightness and contrast of the SAR image are improved through image gray level histogram equalization, so that the display of image detail information is more obvious.
Denoising an image: then, denoising pretreatment is carried out on the SAR image by using an adaptive denoising algorithm based on NSCT, so that speckle noise is removed and image detail texture information is protected.
(2) Preliminary detection of ship position
A deep convolutional neural network-based SAR image ship target detection method becomes a mainstream method in the field, however, in the deep neural network, small target characteristic information is easy to lose, and finally small target detection omission is caused. Aiming at the problem, a network structure (SSS-YOLOv 3) for a small target is designed, a cavity convolution module is introduced to expand a perception field, and preliminary detection of the SAR image ship target is realized.
The ships near the coast are densely arranged, and the prediction frame with a low score in the overlapped frame can be removed by the traditional non-maximum inhibition strategy. The prediction boxes of different targets in a dense area may be erroneously deleted. The conventional non-maximum suppression (NMS) formula is as follows:
(3) ship target accurate detection based on multi-scale segmentation and contour estimation
Multi-scale threshold segmentation: on the basis of primary detection of ships, binary segmentation is carried out on the SAR image by using a global OTSU threshold, a foreground pixel value is 1, a background pixel value is 0, and connectivity area extraction is carried out on the OTSU threshold segmentation result to obtain a foreground connected area and a background connected area. Secondly, performing threshold operation on the area of the foreground communication area, and setting the foreground area smaller than a certain area as a background area to remove the influence of sea clutter. And finally, performing morphological expansion operation on the detection result to obtain a complete ship target area.
Extracting characteristic information: on the basis of obtaining the accurate position of a ship target, a multi-layer feature fusion method is adopted to respectively calculate the detected geometric features (length, width, length-width ratio, area, perimeter and the like), shape features (shape index and the like), edge features and scattering features, so that ship detection is realized.
The invention has the beneficial effects that:
(1) the position, the direction and the contour of the SAR image ship target are detected more accurately.
Aiming at the fact that the traditional ship detection method can only extract the approximate position of a ship on an SAR image, the minimum circumscribed rectangle of a ship target can be obtained by adopting a multi-scale OTSU threshold segmentation and target contour estimation method, and therefore accurate azimuth and position information of the ship is extracted.
(2) And multi-layer characteristic information of the ship target is further extracted.
On the basis of obtaining the accurate position of the ship target, the geometric features (length, width, length-width ratio, area, circumference and the like), shape features (shape index and the like), edge features, scattering features and the like of the detection are respectively calculated by adopting a multi-layer feature fusion method, so that the ship target can be more accurately detected and identified.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
FIG. 2 is a schematic view of detection.
Detailed Description
Referring to fig. 1 and fig. 2, the process and embodiment of the present invention specifically include the following steps:
(1) image pre-processing
Selecting three typical satellite-borne SAR images, and improving the brightness and contrast of the SAR images through image gray level histogram equalization; then, speckle noise of the SAR image is eliminated by using an NSCT-based adaptive denoising algorithm, and image edge information can be reserved to a greater extent.
(2) Preliminary detection of ship position
On the basis of SAR image preprocessing, a deep convolutional neural network model (such as SSS-YOLOv 3) is adopted, and a hole convolution module is introduced to expand a perception field of view, so that preliminary detection of SAR image ship targets is realized;
(3) accurate detection of ship targets
On the basis of primary detection of the ship, the SAR image is segmented by using a multi-scale OTSU threshold segmentation method to obtain the minimum circumscribed rectangle of the ship target, and the position and orientation information of the ship is further extracted based on a contour estimation method. On the basis, the geometric characteristics (length, width, length-width ratio, area, circumference and the like), shape characteristics (shape index and the like), edge characteristics, scattering characteristics and the like of the ship are respectively calculated by adopting a multi-layer characteristic fusion method, so that the ship target can be more accurately detected and identified.
The present invention is not limited to the above-described specific embodiments, and various modifications and variations are possible. Any modifications, equivalents, improvements and the like made to the above embodiments in accordance with the technical spirit of the present invention should be included in the scope of the present invention.
Claims (1)
1. A SAR image ship detection method is characterized in that: the method comprises the following steps:
(1) image pre-processing
Image enhancement: the brightness and the contrast of the SAR image are improved through image gray level histogram equalization, so that the detailed information of the image is displayed more obviously;
denoising an image: then, denoising pretreatment is carried out on the SAR image by using an adaptive denoising algorithm based on NSCT, so that speckle noise is removed and image detail texture information is protected;
(2) preliminary detection of ship position
On the basis of SAR image preprocessing, a deep convolution-based neural network model is adopted, and a hole convolution module is introduced to expand a perception field of view, so that primary detection of SAR image ship targets is realized;
(3) ship target accurate detection based on multi-scale segmentation and contour estimation
Multi-scale threshold segmentation: on the basis of primary detection of ships, binary segmentation is carried out on the SAR image by using a global OTSU threshold, a foreground pixel value is 1, a background pixel value is 0, and connectivity area extraction is carried out on the OTSU threshold segmentation result to obtain a foreground connected area and a background connected area; secondly, performing threshold operation on the area of the foreground communication area, and setting the foreground area smaller than the set area as a background area to remove the influence of sea clutter; finally, performing morphological expansion operation on the detection result to obtain a complete ship target area;
extracting characteristic information: on the basis of obtaining the accurate position of a ship target, the geometric characteristics, the shape characteristics, the edge characteristics and the scattering characteristics of ship detection are respectively calculated by adopting a multilayer characteristic fusion method, so that the ship detection is realized.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111606636.7A CN114283322A (en) | 2021-12-26 | 2021-12-26 | SAR image ship detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111606636.7A CN114283322A (en) | 2021-12-26 | 2021-12-26 | SAR image ship detection method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114283322A true CN114283322A (en) | 2022-04-05 |
Family
ID=80875675
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111606636.7A Pending CN114283322A (en) | 2021-12-26 | 2021-12-26 | SAR image ship detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114283322A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114972918A (en) * | 2022-05-30 | 2022-08-30 | 中国人民解放军国防科技大学 | Remote sensing image ship target identification method based on integrated learning and AIS data |
-
2021
- 2021-12-26 CN CN202111606636.7A patent/CN114283322A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114972918A (en) * | 2022-05-30 | 2022-08-30 | 中国人民解放军国防科技大学 | Remote sensing image ship target identification method based on integrated learning and AIS data |
CN114972918B (en) * | 2022-05-30 | 2024-04-19 | 中国人民解放军国防科技大学 | Remote sensing image ship target identification method based on integrated learning and AIS data |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107145874B (en) | Ship target detection and identification method in complex background SAR image | |
CN109117802B (en) | Ship detection method for large-scene high-resolution remote sensing image | |
CN103942803B (en) | SAR (Synthetic Aperture Radar) image based automatic water area detection method | |
CN108229342B (en) | Automatic sea surface ship target detection method | |
CN112734761B (en) | Industrial product image boundary contour extraction method | |
CN111369570B (en) | Multi-target detection tracking method for video image | |
CN113420658A (en) | SAR image sea-land segmentation method based on FCM clustering and OTSU segmentation | |
CN109829858B (en) | Ship-borne radar image oil spill monitoring method based on local adaptive threshold | |
CN106803078B (en) | SAR image ship target segmentation method | |
CN103971370A (en) | Intelligent ocean oil spill detection method for remote sensing large image | |
CN111353371A (en) | Coastline extraction method based on satellite-borne SAR image | |
CN111259866B (en) | Marine ship target detection method based on improved background difference method | |
CN111105390B (en) | Improved sea-sky-line detection and evaluation method | |
CN114764801A (en) | Weak and small ship target fusion detection method and device based on multi-vision significant features | |
Li et al. | Automatic infrared ship target segmentation based on structure tensor and maximum histogram entropy | |
CN114283322A (en) | SAR image ship detection method | |
CN115409831A (en) | Star point centroid extraction method and system based on optimal background estimation | |
CN115439497A (en) | Infrared image ship target rapid identification method based on improved HOU model | |
CN109872358B (en) | Marine oil film identification method for shipborne radar image based on active contour model | |
CN117115193A (en) | Infrared ship positioning method based on line inhibition | |
CN114219996A (en) | SAR image berthing ship target detection method | |
CN110097569A (en) | Oil tank object detection method based on color Markov Chain conspicuousness model | |
CN115409954A (en) | Dense point cloud map construction method based on ORB feature points | |
CN111667494A (en) | Method for quickly segmenting highlight target of original image of shipborne radar | |
Zhang et al. | Oil tank detection based on linear clustering saliency analysis for synthetic aperture radar images |
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 |