CN106886988B - Linear target detection method and system based on unmanned aerial vehicle remote sensing - Google Patents
Linear target detection method and system based on unmanned aerial vehicle remote sensing Download PDFInfo
- Publication number
- CN106886988B CN106886988B CN201510919643.0A CN201510919643A CN106886988B CN 106886988 B CN106886988 B CN 106886988B CN 201510919643 A CN201510919643 A CN 201510919643A CN 106886988 B CN106886988 B CN 106886988B
- Authority
- CN
- China
- Prior art keywords
- image
- target
- information
- elevation
- edge
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 28
- 238000000605 extraction Methods 0.000 claims abstract description 29
- 238000000034 method Methods 0.000 claims abstract description 18
- 230000010365 information processing Effects 0.000 claims description 12
- 239000011159 matrix material Substances 0.000 claims description 12
- 230000011218 segmentation Effects 0.000 claims description 12
- 238000007781 pre-processing Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000004044 response Effects 0.000 claims description 6
- 238000003702 image correction Methods 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 4
- 230000008030 elimination Effects 0.000 claims description 3
- 238000003379 elimination reaction Methods 0.000 claims description 3
- 230000009286 beneficial effect Effects 0.000 abstract description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention relates to a linear target detection method and system based on unmanned aerial vehicle remote sensing. The linear target detection method comprises the following steps: step a: acquiring an unmanned aerial vehicle sequence image, and extracting feature points in an overlapping area of the sequence image; step b: matching the extracted feature points to generate a stereo relative, and calculating elevation information according to the stereo relative; step c: and identifying the edge information of the target to be detected according to the elevation information and the color information of the image, and extracting the target to be detected according to the edge information. According to the method, a data acquisition scheme is formulated according to target features to obtain a sequence image with high overlapping degree, a steady SIFT algorithm is adopted to carry out feature point matching, and the obtained elevation data and the image features are combined to carry out target edge identification, so that the extraction of the existing target is realized. The invention can greatly save cost and is beneficial to improving the timeliness and the accuracy of the detection of the linear target in the image.
Description
Technical Field
The invention belongs to the technical field of linear target detection, and particularly relates to a linear target detection method and system based on unmanned aerial vehicle remote sensing.
Background
The extraction of the linear target in the remote sensing image has important significance and extensive research, such as a ditch, a river, a road, an electric power line and the like, and in practical application, a satellite image or an aerial image with high resolution is mostly selected according to the size of the linear target to detect and extract the information of the linear target. The method for detecting the linear target of the remote sensing image is more, and generally, the image which is subjected to geometric correction is segmented by adopting the appearance characteristics of a two-dimensional color image, so that the method has certain applicability and limitation; if the Hough transformation is used for detecting the main characteristic straight line of the image, the algorithm is based on full search, the requirements on calculated amount and storage space are great, and the real-time processing is not facilitated; although the methods of linear target detection based on mathematical morphology, linear target detection based on region growth, linear target detection based on neural network and the like are simple in calculation and can realize real-time processing, discontinuous straight lines cannot be processed due to the limitation of an edge detection algorithm to a certain extent. Tracking and detection of high-resolution satellite images and aerial images often require purchasing of a large-range remote sensing image for identification and extraction, which causes unnecessary image purchase cost and burden of large-scale data processing.
Disclosure of Invention
The invention provides a linear target detection method and system based on unmanned aerial vehicle remote sensing, and aims to solve the problems in the prior art at least to a certain extent.
The invention has the following implementation mode, and discloses a linear target detection method based on unmanned aerial vehicle remote sensing, which comprises the following steps:
step a: acquiring an unmanned aerial vehicle sequence image, and extracting feature points in an overlapping area of the sequence image;
step b: matching the extracted feature points to generate a stereo relative, and calculating elevation information according to the stereo relative;
step c: and identifying the edge information of the target to be detected according to the elevation information and the color information of the image, and extracting the target to be detected according to the edge information.
The technical scheme adopted by the embodiment of the invention also comprises the following steps: the step a also comprises the following steps: analyzing the target characteristics of the target to be detected, and setting the required resolution and course/sidewise overlapping degree of the image according to the target characteristics; making a flight plan and a shooting rule according to the resolution and the course/sidewise overlapping degree required by the image; the target characteristics comprise position, size and elevation, the flight plan comprises a flight route, relative ground flight height and speed, and the shooting rule comprises a shooting mode.
The technical scheme adopted by the embodiment of the invention also comprises the following steps: the step a further comprises: preprocessing the acquired sequence image; the preprocessing comprises image positioning, actual overlapping degree analysis and image color enhancement.
The technical scheme adopted by the embodiment of the invention also comprises the following steps: in the step a, the feature point extraction method for extracting feature points in an overlapping area of a sequence image includes:
step a 1: searching local three-dimensional extreme points on the scale image in the image scale space by adopting a DoG operator, and preliminarily determining the position and the characteristic scale of the key point;
step a 2: performing Taylor second-order expression expansion on the DoG operator at the key points, and accurately determining the positions and the characteristic scales of the key points by fitting a second-order Taylor expansion expression in an image scale space;
step a 3: forming a Hessian matrix by the first-order differential and the second-order differential of the DoG operator, and comparing the ratio of the maximum eigenvalue to the minimum eigenvalue of the Hessian matrix with a set threshold value to remove low-contrast key points and unstable edge response points and improve the noise resistance of the key points;
step a 4: and forming a gradient direction histogram by the image point gradient direction in the neighborhood of the key point and the gradient value subjected to Gaussian weighting, fitting a value near the maximum value of the histogram by using a parabola, and accurately determining the main direction of the key point to form the SIFT feature point.
The technical scheme adopted by the embodiment of the invention also comprises the following steps: the step b and the step c further comprise the following steps: and acquiring three-dimensional structure information of the area where the target to be detected is located according to the stereo relative mode, and correcting and inlaying the image according to the three-dimensional structure information of the area.
The technical scheme adopted by the embodiment of the invention also comprises the following steps: in the step c, identifying edge information of the target to be detected according to the elevation information and the color information of the image, and extracting the target to be detected according to the edge information specifically includes the following steps:
step c 1: segmenting the color information of the image by means of a mean shift segmentation algorithm to obtain an image edge probability graph;
step c 2: segmenting the elevation information by means of a mean shift segmentation algorithm to obtain an elevation discontinuous boundary probability map;
step c 3: identifying edge information of the target to be detected according to the image edge probability graph and the elevation discontinuous boundary probability graph;
step c 4: and extracting the target to be detected according to the edge information of the target to be detected.
The embodiment of the invention adopts another technical scheme that: a linear target detection system based on unmanned aerial vehicle remote sensing comprises an image acquisition module, a feature point extraction module, a feature point matching module, an elevation calculation module and a target extraction module; the image acquisition module is used for acquiring unmanned aerial vehicle sequence images; the characteristic point extraction module is used for extracting characteristic points in an overlapping area of the sequence images; the characteristic point matching module is used for matching the extracted characteristic points to generate a stereo relative; the elevation calculation module is used for calculating elevation information according to the stereo relative; the target extraction module is used for identifying the edge information of the target to be detected according to the elevation information and the color information of the image and extracting the target to be detected according to the edge information.
The technical scheme adopted by the embodiment of the invention also comprises the following steps: the system also comprises an image setting module, a plan making module and an image correction module, wherein the image setting module is used for analyzing the target characteristics of the target to be detected and setting the required resolution and the course/collateral overlapping degree of the image according to the target characteristics; the plan making module makes a flight plan and a shooting rule according to the resolution and the course/sidewise overlapping degree of the required images; the image correction module is used for acquiring three-dimensional structure information of an area where the target to be detected is located according to the stereo relative mode and correcting and inlaying the image according to the three-dimensional structure information of the area.
The technical scheme adopted by the embodiment of the invention also comprises the following steps: the feature point extraction module comprises a key point search unit, a key point determination unit, a key point elimination unit and a feature point generation unit;
the key point searching unit is used for searching a local three-dimensional extreme point on the scale image in the image scale space by adopting a DoG operator, and preliminarily determining the position and the characteristic scale of the key point;
the key point determining unit is used for carrying out Taylor second-order expansion on the DoG operator at the key point, and accurately determining the position and the characteristic scale of the key point by fitting a second-order Taylor expansion in an image scale space;
the key point removing unit is used for forming a Hessian matrix by the first-order differential and the second-order differential of the DoG operator, and removing key points with low contrast and unstable edge response points by comparing the ratio of the maximum characteristic value and the minimum characteristic value of the Hessian matrix with a set threshold value, so that the noise resistance of the key points is improved;
the feature point generating unit is used for forming a gradient direction histogram by the image point gradient direction in the neighborhood of the key point and the gradient value subjected to Gaussian weighting processing, fitting a value near the maximum value of the histogram by using a parabola, accurately determining the main direction of the key point and forming the SIFT feature point.
The technical scheme adopted by the embodiment of the invention also comprises the following steps: the target extraction module comprises a color information processing unit, an elevation information processing unit, an edge identification unit and a target extraction unit;
the color information processing unit is used for segmenting the color information of the image through a mean shift segmentation algorithm to obtain an image edge probability graph;
the elevation information processing unit is used for segmenting elevation information through a mean shift segmentation algorithm to obtain an elevation discontinuous boundary probability map;
the edge identification unit is used for identifying the edge information of the target to be detected according to the image edge probability map and the elevation discontinuous boundary probability map;
the target extraction unit is used for extracting the target to be detected according to the edge information of the target to be detected.
The linear target detection method and system based on unmanned aerial vehicle remote sensing provided by the embodiment of the invention can be used for formulating a data acquisition scheme according to target characteristics to obtain a sequence image with high overlapping degree, matching characteristic points by adopting a steady SIFT algorithm, and identifying target edges by combining the obtained elevation data with image characteristics, thereby realizing the extraction of the existing target. The invention can greatly save cost and is beneficial to improving the timeliness and the accuracy of the detection of the linear target in the image.
Drawings
FIG. 1 is a flow chart of a linear target detection method based on unmanned aerial vehicle remote sensing according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating a method for extracting feature points in an image overlap region according to an embodiment of the present invention;
FIG. 3 is a flowchart of a method for extracting a target to be detected according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of a linear target detection system based on unmanned aerial vehicle remote sensing in the embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Fig. 1 is a flowchart of a linear target detection method based on unmanned aerial vehicle remote sensing according to an embodiment of the present invention. The linear target detection method provided by the embodiment of the invention comprises the following steps of:
step 100: analyzing the target characteristics of the target to be detected, and setting the required resolution and course/sidewise overlapping degree of the image according to the target characteristics;
in step 100, the target features include position, size, elevation, and the like; determining the range of a detection area according to the position of the detection area, and determining the optimal resolution required by an image according to the size of a target to be detected, so as to inversely calculate the required camera lens condition and the flight height relative to the ground during flight; in order to ensure that the post-processing has enough overlapping area, consider the instability of the platform caused by the influence of the wind speed and the airflow in the flight process, and ensure enough course/side direction overlapping degree, the embodiment of the invention designs the exposure point position of the camera in the data acquisition stage into a mode of course overlapping degree of 80% and side direction overlapping degree of 60%; in practical application, the setting or the modification can be carried out according to specific situations.
Step 200: making a flight plan and a shooting rule according to the resolution and the course/sidewise overlapping degree required by the image;
in step 200, the flight plan includes a flight route, a relative ground altitude, a relative ground speed and the like, and the shooting rule includes a shooting mode and the like.
Step 300: acquiring an unmanned aerial vehicle sequence image according to a flight plan, and preprocessing the acquired sequence image;
in step 300, the sequential image pre-processing includes image positioning, actual overlap analysis, and image color enhancement.
Step 400: SIFT feature point extraction is carried out in the overlapping area of the sequence images;
in step 400, the method adopts a stable SIFT algorithm to extract the feature points in the image overlapping region, and the SIFT algorithm not only has strong adaptability to the complex geometric deformation of the image, but also shows certain superiority in the aspects of operation speed, positioning accuracy and the like. Fig. 2 is a flowchart illustrating a method for extracting feature points in an image overlap area according to an embodiment of the present invention. The method for extracting the feature points of the image overlapping area comprises the following steps of:
step 401: adopting a DoG (difference of Gaussian) operator to search local three-dimensional extreme points on the scale image in the image scale space, and preliminarily determining the position and the characteristic scale of a key point;
step 402: performing Taylor second-order expression expansion on the DoG operator at the key points, and accurately determining the positions and the characteristic scales of the key points by fitting a second-order Taylor expansion expression in an image scale space;
step 403: forming a Hessian matrix by the first-order differential and the second-order differential of the DoG operator, and comparing the ratio of the maximum eigenvalue to the minimum eigenvalue of the Hessian matrix with a set threshold value to remove low-contrast key points and unstable edge response points and improve the noise resistance of the key points;
step 404: and forming a gradient direction histogram by the image point gradient direction in the neighborhood of the key point and the gradient value subjected to Gaussian weighting, fitting a value near the maximum value of the histogram by using a parabola, accurately determining the main direction of the key point, and finally forming the SIFT feature point.
Step 500: SIFT feature point matching is carried out, and the error matching point pair is deleted to generate stereo relativity;
in step 500, the SIFT feature point matching method is as follows: performing coarse matching on the feature points by adopting a BBF (approximate nearest neighbor search) algorithm, and extracting a matching point pair with higher reliability; then, fine matching is carried out on the feature points by adopting a Hough transform algorithm, and wrong matching point pairs caused by self-similarity or symmetry of the image are eliminated;
step 600: calculating elevation information according to the stereo relative;
step 700: acquiring three-dimensional structure information of an area where a target to be detected is located according to the stereo relative mode, and correcting and embedding the image according to the three-dimensional structure information of the area;
step 800: and identifying the edge information of the target to be detected according to the elevation information and the color information of the image, and extracting the target to be detected according to the edge information.
In step 800, on the basis of the acquired elevation information, the method combines the acquired elevation information with the corrected two-dimensional color image, analyzes the characteristic space by adopting a mean shift algorithm in a mean shift process, and enables each data point to be replaced by a corresponding mode point in the characteristic space through mean shift filtering; fig. 3 is a flowchart of a method for extracting a target to be detected according to an embodiment of the present invention. The method for extracting the target to be detected comprises the following steps:
step 801: segmenting the color information of the image by means of a mean shift segmentation algorithm to obtain an image edge probability graph;
step 802: segmenting the elevation information by means of a mean shift segmentation algorithm to obtain an elevation discontinuous boundary probability map;
step 803: identifying edge information of the target to be detected according to the image edge probability graph and the elevation discontinuous boundary probability graph;
step 804: and extracting the target to be detected according to the edge information of the target to be detected.
Please refer to fig. 4, which is a schematic structural diagram of a linear target detection system based on unmanned aerial vehicle remote sensing according to an embodiment of the present invention. The linear target detection system comprises an image setting module, a plan making module, an image obtaining module, a characteristic point extracting module, a characteristic point matching module, an elevation calculating module, an image correcting module and a target extracting module; wherein,
the image setting module is used for analyzing the target characteristics of the target to be detected and setting the required resolution and the course/lateral overlapping degree of the image according to the target characteristics; wherein the target characteristics comprise position, size, elevation and the like; in the embodiment of the invention, the exposure point position of the camera in the data acquisition stage is designed into a mode with course overlapping degree of 80% and side overlapping degree of 60%; in practical application, the setting or the modification can be carried out according to specific situations.
The plan making module makes a flight plan and a shooting rule according to the resolution and the course/sidewise overlapping degree of the required images; the flight plan comprises a flight route, a relative ground flight height, a relative ground flight speed and the like, and the shooting rule comprises a shooting mode and the like;
the image acquisition module is used for acquiring an unmanned aerial vehicle sequence image according to the flight plan and preprocessing the acquired sequence image; the sequence image preprocessing comprises image positioning, actual overlapping degree analysis, image color enhancement and the like.
The feature point extraction module is used for extracting SIFT feature points in an overlapping area of the sequence images; specifically, the feature point extraction module comprises a key point search unit, a key point determination unit, a key point elimination unit and a feature point generation unit;
the key point searching unit is used for searching a local three-dimensional extreme point on the scale image in the image scale space by adopting a DoG operator, and preliminarily determining the position and the characteristic scale of the key point;
the key point determining unit is used for carrying out Taylor second-order expression expansion on the DoG operator at the key point, and accurately determining the position and the characteristic scale of the key point by fitting a second-order Taylor expansion expression in an image scale space;
the key point removing unit is used for forming a Hessian matrix by the first-order differential and the second-order differential of the DoG operator, and removing the key points with low contrast and unstable edge response points by comparing the ratio of the maximum characteristic value and the minimum characteristic value of the Hessian matrix with a set threshold value, so that the noise resistance of the key points is improved;
the feature point generating unit is used for forming a gradient direction histogram by the image point gradient direction in the neighborhood of the key point and the gradient value subjected to Gaussian weighting processing, fitting a value near the maximum value of the histogram by using a parabola, accurately determining the main direction of the key point and finally forming the SIFT feature point.
The characteristic point matching module is used for carrying out SIFT characteristic point matching and deleting mismatching point pairs to generate stereo relativity; the mode of the feature point matching module for matching SIFT feature points is as follows: performing coarse matching on the feature points by adopting a BBF algorithm, and extracting matching point pairs with higher reliability; then, fine matching is carried out on the feature points by adopting a Hough transform algorithm, and wrong matching point pairs caused by self-similarity or symmetry of the image are eliminated;
the elevation calculation module is used for calculating elevation information according to the stereo relative;
the image correction module is used for acquiring three-dimensional structure information of an area where the target to be detected is located according to the stereo relative mode and correcting and embedding the image according to the three-dimensional structure information of the area;
the target extraction module is used for identifying the edge information of the target to be detected according to the elevation information and the color information of the image and extracting the target to be detected according to the edge information; specifically, the target extraction module comprises a color information processing unit, an elevation information processing unit, an edge identification unit and a target extraction unit;
the color information processing unit is used for segmenting the color information of the image through a mean shift segmentation algorithm to obtain an image edge probability graph;
the elevation information processing unit is used for segmenting elevation information through a mean shift segmentation algorithm to obtain an elevation discontinuous boundary probability map;
the edge identification unit is used for identifying the edge information of the target to be detected according to the image edge probability map and the elevation discontinuous boundary probability map;
the target extraction unit is used for extracting the target to be detected according to the edge information of the target to be detected.
The linear target detection method and system based on unmanned aerial vehicle remote sensing provided by the embodiment of the invention can be used for formulating a data acquisition scheme according to target characteristics to obtain a sequence image with high overlapping degree, matching characteristic points by adopting a steady SIFT algorithm, and identifying target edges by combining the obtained elevation data with image characteristics, thereby realizing the extraction of the existing target. The invention can greatly save cost and is beneficial to improving the timeliness and the accuracy of the detection of the linear target in the image.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (7)
1. A linear target detection method based on unmanned aerial vehicle remote sensing comprises the following steps:
step a: acquiring unmanned aerial vehicle sequence images, extracting characteristic points in an overlapping area of the sequence images, and preprocessing the acquired sequence images; the preprocessing comprises image positioning, actual overlapping degree analysis and image color enhancement;
specifically, step a 1: searching local three-dimensional extreme points on the scale image in the image scale space by adopting a DoG operator, and preliminarily determining the position and the characteristic scale of the key point;
step a 2: performing Taylor second-order expression expansion on the DoG operator at the key points, and accurately determining the positions and the characteristic scales of the key points by fitting a second-order Taylor expansion expression in an image scale space;
step a 3: forming a Hessian matrix by the first-order differential and the second-order differential of the DoG operator, and comparing the ratio of the maximum eigenvalue to the minimum eigenvalue of the Hessian matrix with a set threshold value to remove low-contrast key points and unstable edge response points and improve the noise resistance of the key points;
step a 4: forming a gradient direction histogram by the image point gradient direction in the neighborhood of the key point and the gradient value subjected to Gaussian weighting processing, fitting a value near the maximum value of the histogram by using a parabola, and accurately determining the main direction of the key point to form SIFT feature points;
step b: matching the extracted feature points to generate a stereo relative, and calculating elevation information according to the stereo relative;
step c: and identifying the edge information of the target to be detected according to the elevation information and the color information of the image, and extracting the target to be detected according to the edge information.
2. The method of claim 1, wherein step a is preceded by: analyzing the target characteristics of the target to be detected, and setting the required resolution and course/sidewise overlapping degree of the image according to the target characteristics; making a flight plan and a shooting rule according to the resolution and the course/sidewise overlapping degree required by the image; the target characteristics comprise position, size and elevation, the flight plan comprises a flight route, relative ground flight height and speed, and the shooting rule comprises a shooting mode.
3. The method of claim 1, further comprising, between steps b and c: and acquiring three-dimensional structure information of the area where the target to be detected is located according to the stereo relative mode, and correcting and inlaying the image according to the three-dimensional structure information of the area.
4. The method according to claim 3, wherein in the step c, the identifying the edge information of the target according to the elevation information and the color information of the image, and the extracting the target according to the edge information specifically comprises the following steps:
step c 1: segmenting the color information of the image by means of a mean shift segmentation algorithm to obtain an image edge probability graph;
step c 2: segmenting the elevation information by means of a mean shift segmentation algorithm to obtain an elevation discontinuous boundary probability map;
step c 3: identifying edge information of the target to be detected according to the image edge probability graph and the elevation discontinuous boundary probability graph;
step c 4: and extracting the target to be detected according to the edge information of the target to be detected.
5. A linear target detection system based on unmanned aerial vehicle remote sensing is characterized by comprising an image acquisition module, a feature point extraction module, a feature point matching module, an elevation calculation module and a target extraction module; the image acquisition module is used for acquiring unmanned aerial vehicle sequence images; the characteristic point extraction module is used for extracting characteristic points in an overlapping area of the sequence images; the characteristic point matching module is used for matching the extracted characteristic points to generate a stereo relative; the elevation calculation module is used for calculating elevation information according to the stereo relative; the target extraction module is used for identifying the edge information of the target to be detected according to the elevation information and the color information of the image and extracting the target to be detected according to the edge information;
the feature point extraction module comprises a key point search unit, a key point determination unit, a key point elimination unit and a feature point generation unit;
the key point searching unit is used for searching a local three-dimensional extreme point on the scale image in the image scale space by adopting a DoG operator, and preliminarily determining the position and the characteristic scale of the key point;
the key point determining unit is used for carrying out Taylor second-order expansion on the DoG operator at the key point, and accurately determining the position and the characteristic scale of the key point by fitting a second-order Taylor expansion in an image scale space;
the key point removing unit is used for forming a Hessian matrix by the first-order differential and the second-order differential of the DoG operator, and removing key points with low contrast and unstable edge response points by comparing the ratio of the maximum characteristic value and the minimum characteristic value of the Hessian matrix with a set threshold value, so that the noise resistance of the key points is improved;
the feature point generating unit is used for forming a gradient direction histogram by the image point gradient direction in the neighborhood of the key point and the gradient value subjected to Gaussian weighting processing, fitting a value near the maximum value of the histogram by using a parabola, accurately determining the main direction of the key point and forming the SIFT feature point.
6. The system of claim 5, further comprising an image setting module, a planning module and an image correction module, wherein the image setting module is configured to analyze a target feature of the target to be detected, and set a resolution and a course/lateral overlap degree required by the image according to the target feature; the plan making module makes a flight plan and a shooting rule according to the resolution and the course/sidewise overlapping degree of the required images; the image correction module is used for acquiring three-dimensional structure information of an area where the target to be detected is located according to the stereo relative mode and correcting and inlaying the image according to the three-dimensional structure information of the area.
7. The system of claim 5, wherein the target extraction module includes a color information processing unit, an elevation information processing unit, an edge recognition unit, and a target extraction unit;
the color information processing unit is used for segmenting the color information of the image through a mean shift segmentation algorithm to obtain an image edge probability graph;
the elevation information processing unit is used for segmenting elevation information through a mean shift segmentation algorithm to obtain an elevation discontinuous boundary probability map;
the edge identification unit is used for identifying the edge information of the target to be detected according to the image edge probability map and the elevation discontinuous boundary probability map;
the target extraction unit is used for extracting the target to be detected according to the edge information of the target to be detected.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510919643.0A CN106886988B (en) | 2015-12-11 | 2015-12-11 | Linear target detection method and system based on unmanned aerial vehicle remote sensing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510919643.0A CN106886988B (en) | 2015-12-11 | 2015-12-11 | Linear target detection method and system based on unmanned aerial vehicle remote sensing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106886988A CN106886988A (en) | 2017-06-23 |
| CN106886988B true CN106886988B (en) | 2020-07-24 |
Family
ID=59173743
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510919643.0A Active CN106886988B (en) | 2015-12-11 | 2015-12-11 | Linear target detection method and system based on unmanned aerial vehicle remote sensing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106886988B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109104574A (en) * | 2018-10-09 | 2018-12-28 | 中国科学院深圳先进技术研究院 | A kind of method and system of cradle head camera masking privacy area |
| CN110232389B (en) * | 2019-06-13 | 2022-11-11 | 内蒙古大学 | Stereoscopic vision navigation method based on invariance of green crop feature extraction |
| CN111310624B (en) * | 2020-02-05 | 2023-11-21 | 腾讯科技（深圳）有限公司 | Occlusion recognition method, occlusion recognition device, computer equipment and storage medium |
| CN112258419B (en) * | 2020-11-02 | 2023-08-11 | 无锡艾立德智能科技有限公司 | Method for enhancing image edge information by weighting |
| CN113155293B (en) * | 2021-04-06 | 2022-08-12 | 内蒙古工业大学 | Human body remote sensing temperature measurement monitoring and recognition system based on unmanned aerial vehicle |
| CN116415910A (en) * | 2023-03-27 | 2023-07-11 | 国网山东省电力公司建设公司 | Unmanned aerial vehicle-based power transmission line environment-friendly intelligent checking method |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101127078A (en) * | 2007-09-13 | 2008-02-20 | 北京航空航天大学 | Unmanned machine vision image matching method based on ant colony intelligence |
| CN101916452A (en) * | 2010-07-26 | 2010-12-15 | 中国科学院遥感应用研究所 | Method for automatically stitching unmanned aerial vehicle remote sensing images based on flight control information |
| CN102254319A (en) * | 2011-04-19 | 2011-11-23 | 中科九度(北京)空间信息技术有限责任公司 | Method for carrying out change detection on multi-level segmented remote sensing image |
| CN102999892A (en) * | 2012-12-03 | 2013-03-27 | 东华大学 | Intelligent fusion method for depth images based on area shades and red green blue (RGB) images |
| CN103714549A (en) * | 2013-12-30 | 2014-04-09 | 南京大学 | Stereo image object segmentation method based on rapid local matching |
| CN104867126A (en) * | 2014-02-25 | 2015-08-26 | 西安电子科技大学 | Method for registering synthetic aperture radar image with change area based on point pair constraint and Delaunay |
| CN105139355A (en) * | 2015-08-18 | 2015-12-09 | 山东中金融仕文化科技股份有限公司 | Method for enhancing depth images |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8957783B2 (en) * | 2012-10-23 | 2015-02-17 | Bounce Imaging, Inc. | Remote surveillance system |
-
2015
- 2015-12-11 CN CN201510919643.0A patent/CN106886988B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101127078A (en) * | 2007-09-13 | 2008-02-20 | 北京航空航天大学 | Unmanned machine vision image matching method based on ant colony intelligence |
| CN101916452A (en) * | 2010-07-26 | 2010-12-15 | 中国科学院遥感应用研究所 | Method for automatically stitching unmanned aerial vehicle remote sensing images based on flight control information |
| CN102254319A (en) * | 2011-04-19 | 2011-11-23 | 中科九度(北京)空间信息技术有限责任公司 | Method for carrying out change detection on multi-level segmented remote sensing image |
| CN102999892A (en) * | 2012-12-03 | 2013-03-27 | 东华大学 | Intelligent fusion method for depth images based on area shades and red green blue (RGB) images |
| CN103714549A (en) * | 2013-12-30 | 2014-04-09 | 南京大学 | Stereo image object segmentation method based on rapid local matching |
| CN104867126A (en) * | 2014-02-25 | 2015-08-26 | 西安电子科技大学 | Method for registering synthetic aperture radar image with change area based on point pair constraint and Delaunay |
| CN105139355A (en) * | 2015-08-18 | 2015-12-09 | 山东中金融仕文化科技股份有限公司 | Method for enhancing depth images |
Non-Patent Citations (1)
| Title |
|---|
| 视频序列中目标的多特征融合跟踪技术研究;郭运艳;《中国优秀硕士学位论文全文数据库 信息科技辑》;20140815;正文第7-8页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106886988A (en) | 2017-06-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106886988B (en) | Linear target detection method and system based on unmanned aerial vehicle remote sensing | |
| CN111028277B (en) | SAR and optical remote sensing image registration method based on pseudo-twin convolution neural network | |
| WO2019042232A1 (en) | Fast and robust multimodal remote sensing image matching method and system | |
| CN107481274B (en) | Robust reconstruction method of three-dimensional crop point cloud | |
| Isikdogan et al. | Automatic channel network extraction from remotely sensed images by singularity analysis | |
| CN110569861B (en) | Image matching positioning method based on point feature and contour feature fusion | |
| CN104200461A (en) | Mutual information image selected block and sift (scale-invariant feature transform) characteristic based remote sensing image registration method | |
| CN110222661B (en) | Feature extraction method for moving target identification and tracking | |
| CN101916373A (en) | Road semiautomatic extraction method based on wavelet detection and ridge line tracking | |
| CN106709499A (en) | SIFT image feature point extraction method based on Canny operator and Hilbert-Huang transform | |
| CN106327522A (en) | Infrared small target detection method based on multi-direction morphological filtering complex cloud background | |
| CN103679740B (en) | ROI (Region of Interest) extraction method of ground target of unmanned aerial vehicle | |
| Zhu et al. | Integrating saliency and ResNet for airport detection in large-size remote sensing images | |
| CN114419165A (en) | Camera external parameter correcting method, device, electronic equipment and storage medium | |
| Zhang et al. | Multiple Saliency Features Based Automatic Road Extraction from High‐Resolution Multispectral Satellite Images | |
| Zou et al. | Path voting based pavement crack detection from laser range images | |
| Wei et al. | Detection of lane line based on Robert operator | |
| CN116863357A (en) | Unmanned aerial vehicle remote sensing dyke image calibration and intelligent segmentation change detection method | |
| CN113724277A (en) | Power line detection method based on Radon transformation | |
| CN117197700A (en) | Intelligent unmanned inspection contact net defect identification system | |
| Wu et al. | Research on crack detection algorithm of asphalt pavement | |
| CN116385477A (en) | Tower image registration method based on image segmentation | |
| CN112734745B (en) | Unmanned aerial vehicle thermal infrared image heating pipeline leakage detection method fusing GIS data | |
| CN109685837B (en) | Heterologous remote sensing image registration method based on feature structure similarity | |
| CN114120140A (en) | Method for automatically extracting building height based on satellite image |
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 |