CN112597905A - Unmanned aerial vehicle detection method based on skyline segmentation - Google Patents
Unmanned aerial vehicle detection method based on skyline segmentation Download PDFInfo
- Publication number
- CN112597905A CN112597905A CN202011559805.1A CN202011559805A CN112597905A CN 112597905 A CN112597905 A CN 112597905A CN 202011559805 A CN202011559805 A CN 202011559805A CN 112597905 A CN112597905 A CN 112597905A
- Authority
- CN
- China
- Prior art keywords
- skyline
- unmanned aerial
- aerial vehicle
- image
- detection
- 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 59
- 230000011218 segmentation Effects 0.000 title claims abstract description 7
- 238000005070 sampling Methods 0.000 claims description 16
- 238000000034 method Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 2
- 238000007781 pre-processing Methods 0.000 abstract description 5
- 230000000694 effects Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
-
- 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
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- 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)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The invention provides an unmanned aerial vehicle detection method based on skyline segmentation, which comprises the following steps: identifying skylines in the image by using a skyline detection algorithm, and identifying skyline areas; and identifying the unmanned aerial vehicle target by using a small target detection algorithm in an area outside the skyline of the image. According to the unmanned aerial vehicle target detection method, the influence of the skyline on the unmanned aerial vehicle detection algorithm is eliminated through preprocessing, so that the calculated amount of unmanned aerial vehicle detection can be reduced, and the detection precision of the unmanned aerial vehicle target is improved.
Description
Technical Field
The invention relates to the technical field of photoelectric detection technology and image processing, in particular to an unmanned aerial vehicle detection method based on skyline segmentation
Background
An Unmanned Aerial Vehicle is an aircraft without an onboard pilot, known by the acronym "UAV" (Unmanned Aerial Vehicle), and is typically controlled for flight by an operator remotely or by autonomous software. The multifunctional hanging platform has the characteristics of capability of hanging various devices, convenience, rapidness, simplicity in operation and no limitation of the physiological structure of an operator, so that the multifunctional hanging platform is developed rapidly in recent years. In contrast to conventional manned aircraft, unmanned aerial vehicles were originally used for tasks where humans were reluctant to drive the aircraft to do so in person because of being too dirty or dangerous. Although unmanned aerial vehicles have mainly originated from military applications, their use is rapidly expanding to civilian fields such as scientific research, commercial activities, entertainment industries, and the like, with great potential.
With the development of economy and science and technology in recent years, civil unmanned aerial vehicles in China have been developed at a high speed, a large number of excellent unmanned aerial vehicle enterprises including 'Dajiang' are emerged, and unmanned aerial vehicle flight is gradually changed into a sport that ordinary people can bear expenses from an expensive and inaccessible taste. Along with it, the civil unmanned aerial vehicle flight hand quantity of china is also constantly leaping up, and they are active in each field such as unmanned aerial vehicle operation, unmanned aerial vehicle racing, unmanned aerial vehicle agricultural broadcast medicine, unmanned aerial vehicle mail express delivery, constantly expand the new way of unmanned aerial vehicle civilization. However, with the increasing popularization of civil unmanned aerial vehicles, the potential safety hazard caused by illegal operation of the flight hands of unqualified unmanned aerial vehicles is increasingly deepened, and the black-flying unmanned aerial vehicles cause more and more serious harm to the aspects of airport safety, security of major security responsibility areas, security of crowded places, national secret protection of confidential areas and the like. In this case, unmanned aerial vehicle no-fly zones are established in a plurality of cities such as beijing city in specific areas by using GPS locks and secret areas such as airports, crowded places, military bases and the like.
However, neither administrative ban nor GPS lock can effectively prevent more and more "black flying" drones from disturbing social security. Some "black-flying" drones do not contain GPS modules themselves, and the use of GPS locks to restrict such drones also loses prerequisites. And the accident that many unmanned aerial vehicles wrongly enter the no-fly zone and lead to the incident is also not that the operator is intended, and the unmanned aerial vehicle is out of control all probably to lead to unmanned aerial vehicle's own unstability, the influence of bad weather condition, radio interference. In view of this, it is of practical significance to develop an unmanned aerial vehicle detection algorithm. The unmanned aerial vehicle target under the flight state has small, slow, the characteristics of height of speed, therefore is difficult to distinguish with the ground object background. Compared with large targets such as ships and warships, the detection difficulty is higher due to less available characteristic information. Traditional aerial target detection relies on radar, and good effects are achieved by utilizing Doppler effect and a continuously improved algorithm. But unmanned aerial vehicle does not possess fast, the big characteristics of reflection area, is difficult to carry out radar detection. Therefore, many research institutes have recently used optical properties to detect drones.
Disclosure of Invention
The invention provides an unmanned aerial vehicle detection method based on skyline segmentation, which eliminates the influence of skylines on an unmanned aerial vehicle detection algorithm through preprocessing, thereby greatly reducing the calculated amount of unmanned aerial vehicle detection and improving the detection precision of the unmanned aerial vehicle detection.
In contrast, the invention provides an unmanned aerial vehicle detection method based on skyline segmentation, which comprises the following steps: identifying skylines in the image by using a skyline detection algorithm, and identifying skyline areas; and identifying the unmanned aerial vehicle target by using a small target detection algorithm in an area outside the skyline of the image.
The skyline detection algorithm comprises the following steps: 1) sampling an image to be processed and establishing an integral image; 2) analyzing the variation of the gray gradient of the pixels of the sampling column to find out a possible area of the skyline, wherein the area is a suspicious skyline area; 3) finding out a point on a suspected skyline from the suspicious region, and calling the point as a suspicious point; 4) screening suspicious points of the skyline and removing non-skyline points; 5) connecting the suspicious points of the skyline obtained in the above steps, and judging and combining skylines of a plurality of different areas; 6) and detecting the region outside the skyline in the image by using a small target detection algorithm, thereby completing the unmanned detection.
Wherein, 1) sampling and establishing integral image for the image to be processed, it includes: sampling the image at specified intervals to establish sampling columns, and calculating integral images of the sampling columns for subsequent processing operations.
In order to reduce the influence of noise, a filtering algorithm for eliminating the influence of noise is added in steps 2) and 3).
The invention has better detection capability on the skyline, can reduce the calculated amount of unmanned detection and improve the detection precision.
Drawings
FIG. 1 is a diagram of the effects of the various stages of the skyline detection algorithm.
FIG. 2 is a schematic illustration of a portion of an image of a siftflow dataset.
Fig. 3 is a schematic diagram of the processing of an image containing a drone by the method.
Detailed Description
The existing unmanned aerial vehicle detection algorithm has limited effect when facing a low-altitude flying unmanned aerial vehicle, and the main reason is that a complex skyline boundary and an unmanned aerial vehicle target are difficult to distinguish when a small target detection algorithm is used for detecting the unmanned aerial vehicle. Therefore, a preprocessing is performed before the small target identification is performed, the skyline is identified, and the accuracy of unmanned aerial detection is greatly improved.
The purpose of the invention is: through preprocessing, the influence of the skyline on the unmanned aerial vehicle detection algorithm is eliminated, so that the calculated amount of the unmanned aerial vehicle detection is greatly reduced, and the detection precision of the unmanned aerial vehicle detection is improved.
The technical scheme is as follows:
the general scheme is that a pretreatment is carried out before the detection of the low-altitude flying unmanned aerial vehicle under the complex background is carried out, and then the unmanned aerial vehicle identification is carried out, namely:
1. identifying the antenna in the graph by using an antenna detection algorithm, and marking a non-target label on the antenna;
2. utilize traditional little target detection algorithm discernment unmanned aerial vehicle target, because skyline part has been got rid of to the last step, consequently the detection performance can promote greatly.
Through the operation of preprocessing, the area needing to be detected is greatly reduced, so that the operation amount is reduced, and the detection accuracy rate is far higher than that of a full-image detection algorithm because the area with larger interference is eliminated.
The processing flow of the algorithm can be summarized as the following key points:
1. and selecting a sampling column and establishing an integral image, and establishing the sampling column and calculating a corresponding integral image at certain intervals so as to facilitate later operation.
2. By analyzing the variation of the gray level gradient of the pixels of the sampling column, the possible area of the skyline is found, and the area is called a suspicious skyline area.
3. From the suspicious region, a point on the suspected skyline is found, called suspicious point. Meanwhile, considering the influence of noise, a filtering algorithm for eliminating the influence of noise is added in the steps 2 and 3, so that the method has stronger robustness. The filtering algorithm includes a kalman filter and the like.
4. And suspicious points of the skyline are screened, and non-skyline points are removed by using a certain rule, so that the accuracy rate of identification of the skyline is improved.
5. And connecting the suspicious points of the skyline obtained in the steps, and judging and combining the skyline in different areas by using a certain rule.
6. And detecting an area outside the skyline in the image by using a traditional unmanned aerial vehicle detection algorithm so as to finish unmanned aerial vehicle detection.
The integral image calculation formula in the step 1 is as follows:
in the formula, p represents a sampling interval, m is the serial number of a sampling column, and y ranges from [0, h-1], and represents the number of vertical pixels of an input image.
When the pixel sum of any length in a certain column needs to be calculated, only corresponding subtraction needs to be carried out on the formula, so that the calculation time is greatly shortened, as shown in the following:
FIG. 1 is a diagram of the effects of the various stages of the skyline detection algorithm.
FIG. 3 is a schematic diagram of a processing flow of an image according to the present method. It can be seen that the skyline is marked after the original image is detected by the skyline. At this point, it has been tagged with a "non-target". Therefore, the skyline area with large influence on the detection of the low-altitude unmanned aerial vehicle in the image is eliminated, and the detection success rate of the unmanned aerial vehicle is greatly improved. The technical effects are as follows: by testing 2688 pairs of images (which are all pictures containing skylines in nature) taken from the siftflow data set, the method has better detection capability on the skylines.
FIG. 2 is a schematic illustration of a portion of an image of a siftflow dataset. Meanwhile, some low-altitude unmanned aerial vehicle flying photos are shot, and the unmanned aerial vehicle is detected by the method, so that the unmanned aerial vehicle has good detection capability.
The method has better detection capability on the skyline, can reduce the calculated amount of unmanned detection and improve the detection precision.
Claims (4)
1. An unmanned aerial vehicle detection method based on skyline segmentation is characterized by comprising the following steps:
identifying skylines in the image by using a skyline detection algorithm, and identifying skyline areas;
and identifying the unmanned aerial vehicle target by using a small target detection algorithm in an area outside the skyline of the image.
2. The method of claim 1, wherein the step of the skyline detection algorithm comprises:
1) sampling an image to be processed and establishing an integral image;
2) analyzing the variation of the gray gradient of the pixels of the sampling column to find out a possible area of the skyline, wherein the area is a suspicious skyline area;
3) finding out a point on a suspected skyline from the suspicious region, and calling the point as a suspicious point;
4) screening suspicious points of the skyline and removing non-skyline points;
5) connecting the suspicious points of the skyline obtained in the above steps, and judging and combining skylines of a plurality of different areas;
6) and detecting the region outside the skyline in the image by using a small target detection algorithm, thereby completing the unmanned detection.
3. The method of claim 1, wherein step 1) samples the image to be processed and creates an integral image comprising: sampling the image at specified intervals to establish sampling columns, and calculating integral images of the sampling columns for subsequent processing operations.
4. The method according to claim 1, wherein a filtering algorithm for eliminating the noise influence is added in steps 2) and 3) in order to reduce the noise influence.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011559805.1A CN112597905A (en) | 2020-12-25 | 2020-12-25 | Unmanned aerial vehicle detection method based on skyline segmentation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011559805.1A CN112597905A (en) | 2020-12-25 | 2020-12-25 | Unmanned aerial vehicle detection method based on skyline segmentation |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112597905A true CN112597905A (en) | 2021-04-02 |
Family
ID=75202580
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011559805.1A Pending CN112597905A (en) | 2020-12-25 | 2020-12-25 | Unmanned aerial vehicle detection method based on skyline segmentation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112597905A (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103697855A (en) * | 2014-01-07 | 2014-04-02 | 中国人民解放军国防科学技术大学 | Hull horizontal attitude measurement method based on sea-sky-line detection |
CN103697883A (en) * | 2014-01-07 | 2014-04-02 | 中国人民解放军国防科学技术大学 | Aircraft horizontal attitude determination method based on skyline imaging |
CN106205217A (en) * | 2016-06-24 | 2016-12-07 | 华中科技大学 | Unmanned plane automatic testing method based on machine vision and unmanned plane method of control |
CN109285177A (en) * | 2018-08-24 | 2019-01-29 | 西安建筑科技大学 | A kind of digital city skyline extracting method |
CN110490155A (en) * | 2019-08-23 | 2019-11-22 | 电子科技大学 | A kind of no-fly airspace unmanned plane detection method |
CN110580043A (en) * | 2019-08-12 | 2019-12-17 | 中国科学院声学研究所 | Water surface target avoidance method based on image target identification |
CN111222511A (en) * | 2020-04-13 | 2020-06-02 | 中山大学 | Infrared unmanned aerial vehicle target detection method and system |
-
2020
- 2020-12-25 CN CN202011559805.1A patent/CN112597905A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103697855A (en) * | 2014-01-07 | 2014-04-02 | 中国人民解放军国防科学技术大学 | Hull horizontal attitude measurement method based on sea-sky-line detection |
CN103697883A (en) * | 2014-01-07 | 2014-04-02 | 中国人民解放军国防科学技术大学 | Aircraft horizontal attitude determination method based on skyline imaging |
CN106205217A (en) * | 2016-06-24 | 2016-12-07 | 华中科技大学 | Unmanned plane automatic testing method based on machine vision and unmanned plane method of control |
CN109285177A (en) * | 2018-08-24 | 2019-01-29 | 西安建筑科技大学 | A kind of digital city skyline extracting method |
CN110580043A (en) * | 2019-08-12 | 2019-12-17 | 中国科学院声学研究所 | Water surface target avoidance method based on image target identification |
CN110490155A (en) * | 2019-08-23 | 2019-11-22 | 电子科技大学 | A kind of no-fly airspace unmanned plane detection method |
CN111222511A (en) * | 2020-04-13 | 2020-06-02 | 中山大学 | Infrared unmanned aerial vehicle target detection method and system |
Non-Patent Citations (1)
Title |
---|
仇荣超 等: "前视红外图像中海岸线与海天线的通用检测方法研究" * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109409283B (en) | Method, system and storage medium for tracking and monitoring sea surface ship | |
US20170132468A1 (en) | Systems and methods for object tracking and classification | |
Hu et al. | Detection of unmanned aerial vehicles using a visible camera system | |
CN108108697B (en) | Real-time unmanned aerial vehicle video target detection and tracking method | |
CN108921803B (en) | Defogging method based on millimeter wave and visible light image fusion | |
WO2006135419A2 (en) | A method and system for improved unresolved target detection using multiple frame association | |
Holmes et al. | Adaptive multispectral CFAR detection of land mines | |
Nagarani et al. | Unmanned Aerial vehicle’s runway landing system with efficient target detection by using morphological fusion for military surveillance system | |
CN108257179B (en) | Image processing method | |
Ghosh et al. | AirTrack: Onboard deep learning framework for long-range aircraft detection and tracking | |
CN114139373B (en) | Multi-sensor automatic collaborative management method for unmanned aerial vehicle reverse vehicle | |
Kumar et al. | Detection of runway and obstacles using electro-optical and infrared sensors before landing | |
Mashaly et al. | Efficient sky segmentation approach for small UAV autonomous obstacles avoidance in cluttered environment | |
Balachandran et al. | A novel approach to detect unmanned aerial vehicle using Pix2Pix generative adversarial network | |
CN112418090B (en) | Real-time detection method for infrared weak and small target under sky background | |
Qin et al. | Dense sampling and detail enhancement network: Improved small object detection based on dense sampling and detail enhancement | |
Ang | Research and design of an airfield runway FOD detection system based on WSN | |
CN112597905A (en) | Unmanned aerial vehicle detection method based on skyline segmentation | |
CN112598032A (en) | Multi-task defense model construction method for anti-attack of infrared image | |
Yuksel et al. | Fusion of forward-looking infrared camera and down-looking ground penetrating radar for buried target detection | |
Minwalla et al. | Detection of airborne collision-course targets for sense and avoid on unmanned aircraft systems using machine vision techniques | |
CN114545414A (en) | Track management method for unmanned aerial vehicle anti-collision radar | |
CN114494359A (en) | Small sample moving object detection method based on abnormal optical flow | |
WO2021072996A1 (en) | Embedded airborne infrared image visual navigation positioning method | |
Wainwright et al. | Fusion of morphological images for airborne target detection |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210402 |
|
RJ01 | Rejection of invention patent application after publication |