CN110188601B - Airport remote sensing image detection method based on learning - Google Patents
Airport remote sensing image detection method based on learning Download PDFInfo
- Publication number
- CN110188601B CN110188601B CN201910305556.4A CN201910305556A CN110188601B CN 110188601 B CN110188601 B CN 110188601B CN 201910305556 A CN201910305556 A CN 201910305556A CN 110188601 B CN110188601 B CN 110188601B
- Authority
- CN
- China
- Prior art keywords
- remote sensing
- sensing image
- airport
- detection
- learning
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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
- G06V10/30—Noise filtering
-
- 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
- G06V10/48—Extraction of image or video features by mapping characteristic values of the pattern into a parameter space, e.g. Hough transformation
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (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)
- Life Sciences & Earth Sciences (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a learning-based airport remote sensing image detection method, which comprises the steps of firstly, carrying out Canny edge detection on a remote sensing image, removing noise interference in the remote sensing image and obtaining specific edge information of the remote sensing image; then extracting the initial coordinate value of the longest straight line segment after the remote sensing image edge detection by using Hough transformation, and calculating to obtain the length of the longest straight line; finally, learning classification prediction is carried out by applying a support vector machine, the coordinate value of the starting point of the longest straight-line segment and the longest length value are cascaded to obtain an enhanced characteristic quantity, and the enhanced characteristic quantity is input into the support vector machine for full learning; therefore, the method simplifies the problem of complex remote sensing image classification, eliminates a large amount of interference information in airport remote sensing image detection, utilizes the strong two-classification capability of the support vector machine, does not need to modify parameters too much, is simple and has 96.5 percent of detection accuracy under the same condition.
Description
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to a learning-based airport remote sensing image detection method.
Background
Airport detection plays a key role in airport navigation and military target attack. Meanwhile, the rapid development of the remote sensing technology realizes the acquisition of high-quality remote sensing images. The difference of the remote sensing images of the airport and the information of the non-airport remote sensing images inevitably causes the difference of image characteristic quantity. Because airport runways typically exhibit a strip-like shape with straight edges and relatively uniform edges, with a range of lengths and widths, the edge lines are generally parallel; the remote sensing image airport runway has obvious gray level and is different from the surrounding environment. These features of remote airport sensing images make airport detection possible. In addition, many existing methods for airport target detection are too complex and cumbersome.
The existing research methods have no two ideas, and adopt an optical remote sensing image airport detection method adopting linear feature extraction and an optical remote sensing image airport detection method adopting an image segmentation technology. The former is based on the edge detection of the optical remote sensing image, and then parallel lines are extracted, so that the purpose of positioning the airport is achieved. The method has poor stability because the result of edge detection greatly affects the straight line extraction effect, is easily confused with other irrelevant targets, and causes detection failure. The latter firstly segments the optical remote sensing image to obtain a so-called region of interest (ROI) (region of interest), and then identifies the ROI based on the length of the airport and data information such as a runway. The correct identification of the ROI is related to whether stable airport features can be obtained or not, and is a key factor influencing the airport detection robustness of the optical remote sensing image.
Disclosure of Invention
In order to solve the problems, the invention provides a learning-based airport remote sensing image detection method, which combines Canny edge detection, Hough transformation and support vector machine to realize high-efficiency detection of airport remote sensing images.
A learning-based airport remote sensing image detection method comprises the following steps:
s1: removing noise interference in each remote sensing image by adopting a Canny edge detection algorithm to obtain edge line segment images of each remote sensing image, wherein the remote sensing images comprise airport images and non-airport images;
s2: extracting coordinates (x1, y1, x2, y2) and length z of a starting point of a longest straight line in each edge line segment image by adopting Hough transform to obtain an enhanced characteristic quantity (x1, y1, x2, y2, z) of each remote sensing image, wherein x1 and y1 are respectively an abscissa and an ordinate of a starting point of the longest straight line, and x2 and y2 are respectively an abscissa and an ordinate of a terminal point of the longest straight line;
s3: training a support vector machine by taking the enhanced characteristic quantity corresponding to the airport image as a positive sample and taking the enhanced characteristic quantity corresponding to the non-airport image as a negative sample to obtain an airport target detection model;
s4: re-acquiring the remote sensing image as a detection sample, and then repeatedly executing the steps S1 and S2 to obtain an enhanced characteristic quantity corresponding to the detection sample;
s5: and inputting the enhanced characteristic quantity corresponding to the detection sample into the airport target detection model to realize the detection of the airport target.
Further, when the support vector machine is trained, the adopted kernel function is the RBF radial basis function.
Further, before the Canny edge detection algorithm is adopted to remove noise interference in each remote sensing image, each remote sensing image is converted into a gray level image.
Has the advantages that:
the invention provides a learning-based airport remote sensing image detection method, which comprises the steps of firstly, carrying out Canny edge detection on a remote sensing image, removing noise interference in the remote sensing image and obtaining specific edge information of the remote sensing image; then extracting the initial coordinate value of the longest straight line segment after the edge detection of the remote sensing image by using Hough transform, and calculating to obtain the length of the longest straight line; finally, learning classification prediction is carried out by applying a support vector machine, the coordinate value of the starting point of the longest straight-line segment and the longest length value are cascaded to obtain an enhanced characteristic quantity, and the enhanced characteristic quantity is input into the support vector machine for full learning;
therefore, the noise interference of the remote sensing image is removed in the Canny edge detection link of the remote sensing image, then the proper enhanced characteristic quantity is extracted through Hough transformation, the complicated remote sensing image classification problem is simplified, a large amount of interference information in the airport remote sensing image detection is eliminated, the strong two-classification capability of a support vector machine is utilized, the parameters are not required to be modified too much, and the method is simple and has the detection accuracy rate of 96.5 percent under the same condition.
Drawings
FIG. 1 is a flow chart of a learning-based airport remote sensing image detection method provided by the invention;
FIG. 2 is an original image of an airport remote sensing image provided by the invention;
FIG. 3 is an edge line segment image provided by the present invention;
fig. 4 is a result diagram after hough transform provided by the present invention.
Detailed Description
In order to make the technical solutions better understood by those skilled in the art, 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.
Referring to fig. 1, the figure is a flowchart of a learning-based airport remote sensing image detection method provided in this embodiment. A learning-based airport remote sensing image detection method comprises the following steps:
s1: and removing noise interference in each remote sensing image by adopting a Canny edge detection algorithm to obtain edge line segment images of each remote sensing image, wherein the remote sensing images comprise airport images and non-airport images.
It should be noted that 700 airport remote sensing images and 30800 non-airport remote sensing images can be randomly acquired from Google Earth, wherein the remote sensing image containing the airport target is used as a positive sample image, and the remote sensing image not containing the airport target is used as a negative sample image. As shown in fig. 2, this figure is an original image of the airport remote sensing image according to the present embodiment.
Optionally, before removing noise interference in each remote sensing image by adopting a Canny edge detection algorithm, each remote sensing image is converted into a gray scale image.
It should be noted that Canny edge detection can detect weak edges in images, and even if the contrast between the airport runway and the background is not strong, the detection effect of the operator is good. The main steps of Canny operator for edge detection are as follows:
1) smoothing the image with a gaussian filter eliminates noise.
2) And calculating the amplitude and the direction of the image gradient after filtering.
3) Non-local maximum suppression is applied to the gradient amplitude, and the process is to find out local maximum points in the image gradient and set other non-local maximum points to zero to avoid confusion so as to obtain refined edges.
4) Edges are detected and connected using a hysteresis threshold algorithm.
Referring to fig. 3, the figure is an edge line segment image provided in the present embodiment. As can be seen from FIG. 3, the airport remote sensing image edge information is clear and detailed.
S2: and extracting coordinates (x1, y1, x2 and y2) of a starting point of the longest straight line in each edge line segment image and the length z by adopting Hough transform to obtain the enhanced characteristic quantity (x1, y1, x2, y2 and z) of each remote sensing image, wherein x1 and y1 are respectively an abscissa and an ordinate of the starting point of the longest straight line, and x2 and y2 are respectively an abscissa and an ordinate of the end point of the longest straight line.
It should be noted that, the hough transform extracts the coordinate value of the final start point of the longest straight line segment in the edge line segment image corresponding to the remote sensing image, that is, the hough transform converts the detection problem of the longest straight line segment in the remote sensing image space into the parameter space, completes the detection task by performing simple accumulation statistics in the parameter space, and finally outputs the coordinate value of the final start point of the longest straight line segment, and then calculates the longest length value. And finally, cascading the coordinate value of the final starting point of the longest straight-line segment and the longest length value to obtain the enhanced characteristic quantity.
Referring to fig. 4, the graph is a result graph after hough transform provided in this embodiment. In the figure, the gray-white bright straight line is the longest straight line, and the black line is a confusable straight line.
S3: and training the support vector machine by taking the enhanced characteristic quantity corresponding to the airport image as a positive sample and taking the enhanced characteristic quantity corresponding to the non-airport image as a negative sample to obtain the airport target detection model.
It should be noted that, in this embodiment, 487 airport images in 700 airport images may be used in the positive sample training set, and 195 airport images may be used in the test set, where 18 airport images are not obvious in feature and cannot be extracted; 18512 non-airport images are used for a negative sample training set, 7405 non-airport images are used for a test set, and 4883 non-airport images cannot be used for extracting feature quantities. After the airport target detection model is obtained, the test set is tested to obtain the airport remote sensing image detection accuracy, the powerful two-classification capability of the support vector machine is utilized, parameters do not need to be modified too much, and the airport remote sensing image detection accuracy is 96.5% under the same condition.
S4: re-acquiring the remote sensing image as a detection sample, and then repeatedly executing the steps S1 and S2 to obtain an enhanced characteristic quantity corresponding to the detection sample;
s5: and inputting the enhanced characteristic quantity corresponding to the detection sample into the airport target detection model to realize the detection of the airport target.
Optionally, the embodiment may apply a support vector machine based on the RBF radial basis function to fully learn the training set, so as to generate a detection model of the airport target. And then, whether the input remote sensing image contains an airport target or not can be automatically identified through the detection model.
The present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof, and it will be understood by those skilled in the art that various changes and modifications may be made herein without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (3)
1. A learning-based airport remote sensing image detection method is characterized by comprising the following steps:
s1: removing noise interference in each remote sensing image by adopting a Canny edge detection algorithm to obtain edge line segment images of each remote sensing image, wherein the remote sensing images comprise airport images and non-airport images;
s2: extracting coordinates (x1, y1, x2, y2) and length z of a starting point of a longest straight line in each edge line segment image by adopting Hough transform to obtain an enhanced characteristic quantity (x1, y1, x2, y2, z) of each remote sensing image, wherein x1 and y1 are respectively an abscissa and an ordinate of a starting point of the longest straight line, and x2 and y2 are respectively an abscissa and an ordinate of a terminal point of the longest straight line;
s3: taking the enhanced characteristic quantity corresponding to the airport image as a positive sample, taking the enhanced characteristic quantity corresponding to the non-airport image as a negative sample, and training a support vector machine to obtain an airport target detection model;
s4: re-acquiring the remote sensing image as a detection sample, and then repeatedly executing the steps S1 and S2 to obtain an enhanced characteristic quantity corresponding to the detection sample;
s5: and inputting the enhanced characteristic quantity corresponding to the detection sample into the airport target detection model to realize the detection of the airport target.
2. The airport remote sensing image detection method based on learning of claim 1, wherein a kernel function adopted when training the support vector machine is an RBF radial basis function.
3. The airport remote sensing image detection method based on learning of claim 1, wherein each remote sensing image is converted into a grey-scale map before noise interference in each remote sensing image is removed by a Canny edge detection algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910305556.4A CN110188601B (en) | 2019-04-16 | 2019-04-16 | Airport remote sensing image detection method based on learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910305556.4A CN110188601B (en) | 2019-04-16 | 2019-04-16 | Airport remote sensing image detection method based on learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110188601A CN110188601A (en) | 2019-08-30 |
| CN110188601B true CN110188601B (en) | 2022-07-15 |
Family
ID=67714616
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910305556.4A Active CN110188601B (en) | 2019-04-16 | 2019-04-16 | Airport remote sensing image detection method based on learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110188601B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111161169B (en) * | 2019-12-20 | 2023-06-13 | 五邑大学 | Absolute phase noise removing method and device based on Hough transform and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101777181A (en) * | 2010-01-15 | 2010-07-14 | 西安电子科技大学 | Ridgelet bi-frame system-based SAR image airfield runway extraction method |
| CN102750703A (en) * | 2012-06-27 | 2012-10-24 | 北京航空航天大学 | Remote sensing image airport automatic detecting method based on linear cutting |
| CN103246887A (en) * | 2013-03-26 | 2013-08-14 | 中国科学院电子学研究所 | Airport object multithreading detection method based on optical remote sensing images with geometrical characteristics |
| CN103745224A (en) * | 2013-12-24 | 2014-04-23 | 西南交通大学 | Image-based railway contact net bird-nest abnormal condition detection method |
| CN104156722A (en) * | 2014-08-14 | 2014-11-19 | 西北工业大学 | Airport target detection method based on high-resolution remote sensing image |
| CN105160308A (en) * | 2015-08-20 | 2015-12-16 | 武汉大学 | Airport target recognition method based on line classification and texture classification |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU2018200156B2 (en) * | 2011-05-09 | 2020-06-04 | McVey, Catherine Grace MS | Image Analysis For Determining Characteristics Of Animals And Humans |
| CN104331873A (en) * | 2013-07-22 | 2015-02-04 | 浙江大学 | Method for detecting road from single image |
| CN105373135B (en) * | 2014-08-01 | 2019-01-01 | 深圳中集天达空港设备有限公司 | A kind of method and system of aircraft docking guidance and plane type recognition based on machine vision |
| CN106022375B (en) * | 2016-05-19 | 2019-07-26 | 东华大学 | A kind of clothes fashion recognition methods based on HU not bending moment and support vector machines |
-
2019
- 2019-04-16 CN CN201910305556.4A patent/CN110188601B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101777181A (en) * | 2010-01-15 | 2010-07-14 | 西安电子科技大学 | Ridgelet bi-frame system-based SAR image airfield runway extraction method |
| CN102750703A (en) * | 2012-06-27 | 2012-10-24 | 北京航空航天大学 | Remote sensing image airport automatic detecting method based on linear cutting |
| CN103246887A (en) * | 2013-03-26 | 2013-08-14 | 中国科学院电子学研究所 | Airport object multithreading detection method based on optical remote sensing images with geometrical characteristics |
| CN103745224A (en) * | 2013-12-24 | 2014-04-23 | 西南交通大学 | Image-based railway contact net bird-nest abnormal condition detection method |
| CN104156722A (en) * | 2014-08-14 | 2014-11-19 | 西北工业大学 | Airport target detection method based on high-resolution remote sensing image |
| CN105160308A (en) * | 2015-08-20 | 2015-12-16 | 武汉大学 | Airport target recognition method based on line classification and texture classification |
Non-Patent Citations (5)
| Title |
|---|
| Airport Detection Base on Support Vector Machine from A Single Image;Yanyun Qu 等，;《2005 5th International Conference on Information Communications & Signal Processing》;20060911;第2006年卷;第546-549页 * |
| SAR图像中机场跑道的自动识别研究;杨顺辽 等，;《武汉理工大学学报(交通科学与工程版)》;20060228;第30卷(第01期);第57-58页 * |
| 几何和图论特征对高分辨率遥感影像土地利用分类的影响;万幼川 等，;《武汉大学学报(信息科学版)》;20090705;第34卷(第07期);第794-798页 * |
| 基于支持向量机的机场检测算法;曲延云 等，;《西安交通大学学报》;20060620;第40卷(第06期);第709-713页 * |
| 星载SAR图像机场兴趣区检测算法;张绍明 等，;《计算机工程》;20080831;第34卷;第19-21页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110188601A (en) | 2019-08-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111008961B (en) | Transmission line equipment defect detection method and system, equipment and medium thereof | |
| CN111145209A (en) | Medical image segmentation method, device, equipment and storage medium | |
| CN108921813B (en) | Unmanned aerial vehicle detection bridge structure crack identification method based on machine vision | |
| CN113076802B (en) | Transformer substation switch on-off state image identification method based on lack of disconnected image sample | |
| CN108647597B (en) | Wrist identification method, gesture identification method and device and electronic equipment | |
| CN104615972B (en) | Intelligent identification method and device for pointer instrument | |
| CN112734729B (en) | Water gauge water level line image detection method and device suitable for night light supplement condition and storage medium | |
| CN110245600B (en) | Unmanned aerial vehicle road detection method for self-adaptive initial quick stroke width | |
| CN107274452B (en) | Automatic detection method for acne | |
| CN115272306B (en) | Solar cell panel grid line enhancement method utilizing gradient operation | |
| CN113780110A (en) | Method and device for detecting weak and small targets in image sequence in real time | |
| CN112634288A (en) | Equipment area image segmentation method and device | |
| CN114444565A (en) | Image tampering detection method, terminal device and storage medium | |
| CN106778777B (en) | Vehicle matching method and system | |
| Wei et al. | Detection of lane line based on Robert operator | |
| CN110188601B (en) | Airport remote sensing image detection method based on learning | |
| CN113408519A (en) | Method and system for reading pointer instrument based on template rotation matching | |
| CN113673515A (en) | Computer vision target detection algorithm | |
| CN107710229B (en) | Method, device and equipment for recognizing shape in image and computer storage medium | |
| CN116648723A (en) | Method and device for analyzing microstructure of material | |
| CN105354823A (en) | Tree-ring image edge extraction and segmentation system | |
| CN115761606A (en) | Box electric energy meter identification method and device based on image processing | |
| CN109815791B (en) | Blood vessel-based identity recognition method and device | |
| CN103778629A (en) | Background model real-time updating method for non-coherent radar image | |
| CN110287786B (en) | Vehicle information identification method and device based on artificial intelligence anti-interference |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |