CN113239752B - Unmanned aerial vehicle aerial image automatic identification system - Google Patents
Unmanned aerial vehicle aerial image automatic identification system Download PDFInfo
- Publication number
- CN113239752B CN113239752B CN202110461261.3A CN202110461261A CN113239752B CN 113239752 B CN113239752 B CN 113239752B CN 202110461261 A CN202110461261 A CN 202110461261A CN 113239752 B CN113239752 B CN 113239752B
- Authority
- CN
- China
- Prior art keywords
- image
- edge
- module
- result
- loop
- 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
- 238000002372 labelling Methods 0.000 claims abstract description 37
- 238000004458 analytical method Methods 0.000 claims abstract description 15
- 238000012986 modification Methods 0.000 claims abstract description 13
- 230000004048 modification Effects 0.000 claims abstract description 13
- 230000008676 import Effects 0.000 claims abstract description 10
- 238000010586 diagram Methods 0.000 claims description 28
- 238000001914 filtration Methods 0.000 claims description 18
- 238000004422 calculation algorithm Methods 0.000 claims description 16
- 238000012216 screening Methods 0.000 claims description 11
- 238000001514 detection method Methods 0.000 claims description 9
- 238000003860 storage Methods 0.000 claims description 9
- 238000003708 edge detection Methods 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 7
- 238000007619 statistical method Methods 0.000 claims description 7
- 238000005520 cutting process Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 abstract description 8
- 239000003208 petroleum Substances 0.000 description 8
- 238000007689 inspection Methods 0.000 description 5
- 238000000034 method Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 239000000284 extract Substances 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000004904 shortening Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to the technical field of image processing, and particularly provides an automatic unmanned aerial vehicle aerial image recognition system which comprises a file import module, an image file recognition module, a recognition result labeling module, a labeling result modification module and a result analysis and export module, and solves the problems that the follow-up analysis and processing of the existing unmanned aerial vehicle aerial image is manually recognized, the recognition automation level is low, the manual recognition labeling efficiency is low, the required time and the cost are high.
Description
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to an automatic recognition system for aerial images of an unmanned aerial vehicle.
Background
In recent years, unmanned aerial vehicles are increasingly widely applied to the industries of inspection, surveying and mapping and the like, and the unmanned aerial vehicles make the operations become more intelligent and automatic by virtue of the advantages of simplicity and convenience in operation, low cost, high flexibility, strong adaptability and the like.
Unmanned aerial vehicle inspection and mapping generally require that an image pickup device is carried on the unmanned aerial vehicle, pictures of a plurality of target areas are continuously shot, information such as positions and angles of image shooting is recorded, two-dimensional or three-dimensional images of the target areas are synthesized, then related analysis is carried out, key data in the images are collected and marked, and finally image data are archived.
The technology of image synthesis is relatively mature and can be completely finished by means of a computer, and software such as Pix4Dmapper can automatically synthesize two-dimensional images of a target area with higher quality.
However, the related research and application in the subsequent analysis processing of the image are less, such as labeling residential areas, hospitals, schools, roads, various key facilities and the like, and the recognition is required to be carried out manually, so that the automation level of the recognition is low. The efficiency of manual identification marking is low, and the time and cost are high.
Disclosure of Invention
The invention provides an unmanned aerial vehicle aerial image automatic identification system, which aims to solve the problems that in the prior art, unmanned aerial vehicle aerial image follow-up analysis and processing is manually identified, the automation level of identification is low, the efficiency of manual identification marking is low, and the required time and cost are high.
The invention provides an unmanned aerial vehicle aerial image automatic identification system, which comprises a file import module, an image file identification module, an identification result labeling module, a labeling result modification module, a result analysis and export module, wherein the file import module imports an image selected by a user into the image file identification module; the image file identification module identifies the imported image and transmits the identification result to the identification result labeling module; the identification result labeling module labels the identification result on the image selected by the user; the labeling result modification module modifies the problematic identification result; and the result analysis and export module performs statistical analysis and storage on the identification result.
The file importing module imports the image selected by the user comprises the following steps: and reading the image selected by the user into a memory, decoding and uniformly converting the image into a two-dimensional lattice image in an RGB format for storage.
The image file identification module identifies the imported image, and comprises the following steps:
1) Preprocessing the imported image, wherein the preprocessing comprises the following steps: firstly, converting an imported image into a gray image, and removing noise points on the gray image; then carrying out edge detection and edge refinement on the gray level image with noise removed to obtain a skeleton diagram; finally, connecting the intervals in the skeleton diagram;
2) Taking the skeleton diagram in the step 1) as a plane diagram, taking pixel points with a plurality of branches in the skeleton diagram as nodes, and taking a path between two nodes as an edge;
3) Detecting all the minimum loops in the skeleton diagram of the step 2);
4) Screening all the smallest loops in the step 3), and filtering all the non-rectangular loops;
5) Cutting the imported image according to the residual loop after filtering in the step 4), and classifying after cutting to form an initial recognition result.
The edge detection determines an edge by detecting a rate of change of adjacent portions of the image.
The edge refinement is to refine the detected edge and refine the same line to a single pixel width.
The detection minimum loop adopts an algorithm, and the algorithm comprises the following steps: two nodes connected for all edgesAnd/>Defining an order, forward traversal as/>Reverse traversal is/>; Two count values are reserved for each edge: /(I)And/>Wherein/>Indicating whether this edge was traversed forward,/>Indicating whether the edge was traversed in the reverse direction; in the initial state, for all edges/>，/>; Then, a meeting/>, is arbitrarily selectedOr/>If it isThe edge is traversed forward, otherwise there is/>Traversing the edge in reverse at this time until the edge is returned; if the edge is traversed in the forward direction, the initial current node is set as/>; Otherwise, setting the initial current node as/>; During traversal, each time the next edge in the clockwise direction of the current node is selected/>Traversing and setting/>, according to the traversing directionOr/>; After returning to the initial node, traversing is completed; after the traversing is completed, a minimum loop is found out; if the edge satisfaction/>, still exists in the graph after the traversal is completedOr/>The above steps continue until all edges meet/>And/>。
The algorithm additionally detects a loop which does not belong to the minimum loop, and after the loop which does not belong to the minimum loop is removed after the detection is finished, the rest is all the minimum loops in the skeleton diagram.
And step 4, screening and filtering by calculating the similarity between the loop and the rectangle, wherein the screening and filtering comprises the following steps: firstly enumerating the direction of a rectangle, then adopting a binary search mode to find out the positions of the edges of the rectangle, uniformly taking n points in each edge, calculating the ratio of the distance between the nearest point on the loop perpendicular to the line segment direction and the point and the length of the corresponding edge for each point, finally obtaining the dissimilarity between the loop and the rectangle, and filtering out loops with dissimilarity greater than a preset threshold value, namely filtering out all non-rectangular loops.
The calculation formula of the dissimilarity is as follows，/>For dissimilarity,/>Direction of rectangle,/>The angle of (2) is 0-360 DEG,/>For the dot/>For the corresponding edge length, r i is the distance from each point on the loop closest to that point perpendicular to the line segment direction.
When the labeling result modification module modifies the problematic recognition result, an operator checks whether the labeling of the corresponding region is wrong, and performs the operations of adding/deleting/modifying in the corresponding region to form a final labeling result, namely a final recognition result.
The invention has the beneficial effects that: the unmanned aerial vehicle aerial image automatic identification system provided by the invention has the advantages that the file importing module has a file importing function, and the image selected by the user is imported into the image file identification module; the image file identification module has an image file identification function, extracts, screens and classifies the image files imported by the user in a key area to form a preliminary classification result (identification result), and the identification result labeling module labels the identification result on the image selected by the user; the labeling result modification module modifies the problematic identification result; the result analysis and export module performs statistical analysis and storage on the recognition result, provides reference for application of related industries, and the unmanned aerial vehicle aerial image automatic recognition system can automatically process aerial image images, automatically find out key areas (residence/office building/factory building/hospital and the like) in the images and mark the key areas on the original images, thereby improving the automation level of application of related industries (such as recognition of high-consequence areas of petroleum pipelines), reducing the workload, improving the recognition marking efficiency, shortening the time and reducing the cost.
Drawings
The present invention will be described in further detail with reference to the accompanying drawings.
FIG. 1 is a flow chart of an automated aerial image recognition system for an unmanned aerial vehicle;
FIG. 2 is a schematic diagram of image skeleton optimization;
FIG. 3 is a definition of nodes and edges of an image skeleton;
FIG. 4 is a minimum loop schematic;
FIG. 5 is a schematic diagram of the distances of points on a rectangle to corresponding points on a loop;
FIG. 6 is an exemplary aerial image of an embodiment presented;
FIG. 7 is a graph of the pretreatment result of FIG. 6;
FIG. 8 is a graph of the loop extraction results of FIG. 7;
Fig. 9 is a preliminary recognition result marked on the original image of fig. 6.
Detailed Description
Example 1:
as shown in fig. 1, an unmanned aerial vehicle aerial image automatic recognition system comprises a file importing module, an image file recognition module, a recognition result labeling module, a labeling result modification module, a result analysis and export module, wherein the file importing module imports an image selected by a user into the image file recognition module; the image file identification module identifies the imported image and transmits the identification result to the identification result labeling module; the identification result labeling module labels the identification result on the image selected by the user; the labeling result modification module modifies the problematic identification result; and the result analysis and export module performs statistical analysis and storage on the identification result.
The invention provides an automatic recognition system for unmanned aerial vehicle aerial images, which comprises a front-end display and a back-end processing algorithm, wherein the front-end display comprises a file importing module, a recognition result labeling module, a labeling result modifying module and a result analyzing and exporting module, and is responsible for providing a user interface, facilitating the operation of a user, displaying the processing result and allowing the user to manually correct the result; the back-end processing algorithm comprises an image file identification module and provides a core algorithm support for the system;
the file importing module has an image file importing function, a user selects an image file, and the image file is imported into the image file identifying module to support various image formats such as jpg\bmp\gif\png; the image file identification module has an image file identification function, extracts, screens and classifies the image files imported by the user in a key area to form a preliminary classification result (identification result), and the identification result labeling module labels the identification result on the image selected by the user; the labeling result modification module modifies the problematic identification result, and allows a user to manually adjust the problematic identification result during modification; the result analysis and export module performs statistical analysis and storage on the recognition result, provides reference for application of related industries, and the unmanned aerial vehicle aerial image automatic recognition system can automatically process aerial image images, automatically find out key areas (residence/office building/factory building/hospital and the like) in the images and mark the key areas on the original images, thereby improving the automation level of application of related industries (such as recognition of high-consequence areas of petroleum pipelines), reducing the workload, improving the recognition marking efficiency, shortening the time and reducing the cost.
Example 2:
On the basis of embodiment 1, further, the file importing module imports the image selected by the user includes the following steps: and reading the image selected by the user into a memory, decoding and uniformly converting the image into a two-dimensional lattice image in an RGB format for storage. The method supports various image formats such as jpg\bmp\gif\png, and the like, has strong adaptability, and the part is mainly participated by the front end.
Further, the image file identification module identifies the imported image, which includes the following steps:
1) Preprocessing the imported image, wherein the preprocessing comprises the following steps: firstly, converting an imported image into a gray image, and removing noise points on the gray image; then carrying out edge detection and edge refinement on the gray level image with noise removed to obtain a skeleton diagram; finally, connecting the intervals in the skeleton diagram; the separation (shown in figure 2) in the connection skeleton diagram after the skeleton diagram is obtained is to optimize the skeleton diagram;
2) Taking the skeleton diagram in the step 1) as a plane diagram, taking pixel points with a plurality of branches in the skeleton diagram as nodes, and taking a path between two nodes as edges (shown in fig. 3); in the graph thus built, the degree of each node (Number of edges connected to the node) satisfies: /(I)；
3) Detecting all the minimum loops in the skeleton diagram of the step 2); the minimum loop is a loop which cannot be split into two or more smaller loops (as shown in fig. 4, there are 5 minimum loops in the figure);
4) Screening all the smallest loops in the step 3), and filtering all the non-rectangular loops;
5) Cutting the imported image according to the residual loop after filtering in the step 4), and classifying after cutting to form an initial recognition result.
Further, the edge detection determines an edge by detecting a rate of change of adjacent portions of the image; the algorithm of edge detection is a well-known technique, and a part with a large rate of change is selected as an edge during detection. The method has the advantages of simple and clear edge determination and high accuracy.
Further, the edge refinement is to refine the detected edge and refine the same line to a single pixel width. The obtained image skeleton is more accurate.
Further, the detection minimum loop adopts an algorithm, and the algorithm comprises the following steps: two nodes connected for all edgesAnd/>Defining an order, forward traversal as/>Reverse traversal is/>; Two count values are reserved for each edge: /(I)And/>Wherein/>Indicating whether this edge was traversed forward,/>Indicating whether the edge was traversed in the reverse direction; in the initial state, for all edges/>，/>; Then, a meeting/>, is arbitrarily selectedOr/>If it/>The edge is traversed forward, otherwise there is/>Traversing the edge in reverse at this time until the edge is returned; if the edge is traversed in the forward direction, the initial current node is set as/>; Otherwise, setting the initial current node as/>; During traversal, each time the next edge in the clockwise direction of the current node is selected/>Traversing and setting/>, according to the traversing directionOr (b); After returning to the initial node, traversing is completed; after the traversing is completed, a minimum loop is found out; if the edge satisfaction/>, still exists in the graph after the traversal is completedOr/>The above steps continue until all edges meet/>And/>。
Furthermore, the algorithm additionally detects a loop which does not belong to the minimum loop, and after the loop which does not belong to the minimum loop is removed after the detection is finished, the rest is all the minimum loops in the skeleton diagram. The detection algorithm additionally detects a loop other than the minimum loop, which is a loop around the outside of the whole pattern (as in fig. 4Loops), so after the detection algorithm is performed, the loops with the largest enclosed area need to be removed, and all the smallest loops remain.
Further, the step 4 performs screening filtration by calculating the similarity between the loop and the rectangle, and the screening filtration includes the following steps: firstly enumerating the direction of a rectangle, then adopting a binary search mode to find the positions of the sides of the rectangle, uniformly taking n points (the rectangle has 4n points in total) for each side, calculating the ratio of the distance r i between the nearest point on the loop which is perpendicular to the line segment direction and is closest to the point and the length of the corresponding side for each point, finally obtaining the dissimilarity between the loop and the rectangle, and filtering out loops with the dissimilarity greater than a preset threshold value, namely, filtering out all non-rectangular loops.
Further, the calculation formula of the dissimilarity is as follows
，For dissimilarity,/>Direction of rectangle,/>The angle of (2) is 0-360 DEG,/>For the dot/>For the corresponding edge length, r i is the distance from each point on the loop closest to that point perpendicular to the line segment direction. The algorithm has high accuracy.
Further, when the labeling result modification module modifies the problematic recognition result, the operator checks whether the labeling of the corresponding region is wrong, and performs the adding/deleting/modifying operation on the corresponding region to form a final labeling result, namely a final recognition result.
The software displays the original aerial image on an interface as a bottom plate; then, the preliminary identification result is overlapped on the bottom plate for marking, and the preliminary identification result comprises the range of all the key areas and the classification result of the key areas; the operator can check whether the labeling of the corresponding region is wrong or not, and directly perform the operations of adding/deleting/modifying on the corresponding region on the interface, thereby adjusting the result and forming the final labeling result. Allowing the operator to adjust the labeling results to obtain more accurate results.
Further, the result analysis and export module performs statistical analysis and storage on the identification result. The specific implementation is that according to the final recognition result, the relevant information is counted, and some additional items can be calculated according to a self-defined formula provided by an operator, and finally written into an Excel table designated by a user for reference. The automatic identification system can automatically count, analyze and identify results (such as the number of buildings, the estimated population, the population nearby inflammable and explosive substances (such as petroleum pipelines), and the like) and derive Excel reports, so that workers in related industries can more conveniently conduct data arrangement work, and references are provided for industrial application.
Example 3:
on the basis of the embodiment 2, the following description is given to the application of the unmanned aerial vehicle aerial image automatic identification system in the petroleum pipeline inspection field by combining with the attached drawings:
Fig. 6 is an exemplary view of aerial images of a petroleum pipeline inspection.
Firstly, importing an aerial image picture (figure 6) of petroleum pipeline inspection into an identification system, and preprocessing the imported picture by the identification system to obtain a skeleton diagram (figure 7); then detecting all the minimum loops in the skeleton diagram, screening all the detected minimum loops, and screening out loops which do not accord with the rule, wherein the result is shown in figure 8;
The system will then identify the corresponding region classifier in the original image, and the identification result will be marked on the original image (fig. 9); fig. 9 shows that all buildings are identified and that the roof building is classified as a factory building; however, there are still some misidentification, and 3 smaller non-buildings in the figure are identified as buildings, and at this time, adjustments are needed manually on the software interface, and 3 misidentified areas are deleted.
Finally, the system can carry out statistics and analysis on the final result to generate an Excel table form for subsequent statistical analysis. The table shows the number of buildings near the petroleum pipeline, the number of key buildings (the counts of hospitals, schools and gas stations) and the estimated population (the calculation formula is customized by operators, the system can automatically calculate according to the formula), and the system can provide assistance for the identification of the high-consequence areas of the petroleum pipeline.
In the description of the present invention, it should be understood that, if any, the orientation or positional relationship indicated by the terms or the like is based on the orientation or positional relationship shown in the drawings, rather than indicating or implying that the apparatus or element in question must have a particular orientation, be constructed and operated in a particular orientation, and therefore the terms describing positional relationship in the drawings are merely for illustrative purposes and are not to be construed as limiting the present invention.
The foregoing examples are merely illustrative of the present invention and are not intended to limit the scope of the present invention, and all designs that are the same or similar to the present invention are within the scope of the present invention.
Claims (8)
1. An unmanned aerial vehicle image automatic identification system that takes photo by plane which characterized in that: comprises a file import module, an image file identification module, an identification result labeling module, a labeling result modification module and a result analysis and export module,
The file importing module imports the image selected by the user into the image file identifying module;
the image file identification module identifies the imported image and transmits the identification result to the identification result labeling module;
the image file identification module identifies the imported image, and comprises the following steps:
1) Preprocessing the imported image, wherein the preprocessing comprises the following steps: firstly, converting an imported image into a gray image, and removing noise points on the gray image; then carrying out edge detection and edge refinement on the gray level image with noise removed to obtain a skeleton diagram; finally, connecting the intervals in the skeleton diagram;
2) Taking the skeleton diagram in the step 1) as a plane diagram, taking pixel points with a plurality of branches in the skeleton diagram as nodes, and taking a path between two nodes as an edge;
3) Detecting all the minimum loops in the skeleton diagram of the step 2);
4) Screening all the smallest loops in the step 3), and filtering all the non-rectangular loops; and step 4, screening and filtering by calculating the similarity between the loop and the rectangle, wherein the screening and filtering comprises the following steps: firstly enumerating the directions of the rectangles, then finding out the positions of the edges of the rectangles by adopting a binary search mode, and uniformly taking each edge Calculating the ratio of the distance between the nearest point on the loop perpendicular to the line segment direction and the point to the length of the corresponding edge for each point to finally obtain the dissimilarity between the loop and the rectangle, and filtering out the loops with dissimilarity greater than the preset threshold value, namely filtering out all the non-rectangle loops;
5) Cutting the imported image according to the residual loop after filtering in the step 4), and classifying after cutting to form an initial recognition result;
The identification result labeling module labels the identification result on the image selected by the user;
The labeling result modification module modifies the problematic identification result;
And the result analysis and export module performs statistical analysis and storage on the identification result.
2. The automated unmanned aerial vehicle aerial image recognition system of claim 1, wherein: the file importing module imports the image selected by the user comprises the following steps: and reading the image selected by the user into a memory, decoding and uniformly converting the image into a two-dimensional lattice image in an RGB format for storage.
3. The automated unmanned aerial vehicle aerial image recognition system of claim 1, wherein: the edge detection determines an edge by detecting a rate of change of adjacent portions of the image.
4. The automated unmanned aerial vehicle aerial image recognition system of claim 1, wherein: the edge refinement is to refine the detected edge and refine the same line to a single pixel width.
5. The automated unmanned aerial vehicle aerial image recognition system of claim 1, wherein: the detection minimum loop adopts an algorithm, and the algorithm comprises the following steps: two nodes connected for all edgesAnd/>Defining an order, forward traversal as/>Reverse traversal is/>; Two count values are reserved for each edge: /(I)And/>Wherein/>Indicating whether this edge was traversed forward,/>Indicating whether the edge was traversed in the reverse direction; in the initial state, for all edges/>，/>; Then, a meeting/>, is arbitrarily selectedOr/>If it/>The edge is traversed forward, otherwise there is/>Traversing the edge in reverse at this time until the edge is returned; if the edge is traversed in the forward direction, the initial current node is set as/>; Otherwise, setting the initial current node as/>; During traversal, each time the next edge in the clockwise direction of the current node is selected/>Traversing and setting/>, according to the traversing directionOr/>; After returning to the initial node, traversing is completed; after the traversing is completed, a minimum loop is found out; if the edge satisfaction/>, still exists in the graph after the traversal is completedOr/>The above steps continue until all edges meet/>And/>。
6. The automated unmanned aerial vehicle aerial image recognition system of claim 5, wherein: the algorithm additionally detects a loop which does not belong to the minimum loop, and after the loop which does not belong to the minimum loop is removed after the detection is finished, the rest is all the minimum loops in the skeleton diagram.
7. The automated unmanned aerial vehicle aerial image recognition system of claim 1, wherein: the calculation formula of the dissimilarity is as followsNs is dissimilarity,/>Direction of rectangle,/>The angle of (2) is 0-360 DEG,/>For the dot/>For the corresponding edge length, r i is the distance from each point on the loop closest to that point perpendicular to the line segment direction.
8. The automated unmanned aerial vehicle aerial image recognition system of claim 1, wherein: when the labeling result modification module modifies the problematic recognition result, an operator checks whether the labeling of the corresponding region is wrong, and performs the operations of adding/deleting/modifying in the corresponding region to form a final labeling result, namely a final recognition result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110461261.3A CN113239752B (en) | 2021-04-27 | 2021-04-27 | Unmanned aerial vehicle aerial image automatic identification system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110461261.3A CN113239752B (en) | 2021-04-27 | 2021-04-27 | Unmanned aerial vehicle aerial image automatic identification system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113239752A CN113239752A (en) | 2021-08-10 |
| CN113239752B true CN113239752B (en) | 2024-06-18 |
Family
ID=77129483
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110461261.3A Active CN113239752B (en) | 2021-04-27 | 2021-04-27 | Unmanned aerial vehicle aerial image automatic identification system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113239752B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101567046A (en) * | 2009-06-11 | 2009-10-28 | 北京航空航天大学 | Target recognition method of unmanned aerial vehicle based on minimum circle-cover matching |
| CN108761237A (en) * | 2018-05-29 | 2018-11-06 | 福州大学 | Unmanned plane electric inspection process image vital electrical component diagnoses automatically and labeling system |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8600124B2 (en) * | 2004-09-16 | 2013-12-03 | Imatx, Inc. | System and method of predicting future fractures |
| US8401222B2 (en) * | 2009-05-22 | 2013-03-19 | Pictometry International Corp. | System and process for roof measurement using aerial imagery |
| US9477904B2 (en) * | 2014-01-13 | 2016-10-25 | Here Global B.V. | Systems and methods for refining building alignment in an aerial image |
| CN105825212A (en) * | 2016-02-18 | 2016-08-03 | 江西洪都航空工业集团有限责任公司 | Distributed license plate recognition method based on Hadoop |
| CN106096497B (en) * | 2016-05-28 | 2019-08-06 | 安徽省（水利部淮河水利委员会）水利科学研究院（安徽省水利工程质量检测中心站） | A kind of house vectorization method for polynary remotely-sensed data |
| CN106570468B (en) * | 2016-10-25 | 2019-11-22 | 中国人民解放军空军工程大学 | A Method for Reconstructing Building Outlines from LiDAR Raw Point Clouds |
| CN106650663B (en) * | 2016-12-21 | 2019-07-16 | 中南大学 | Judgment method for true and false changes of buildings and method for removing false changes with this method |
| CN112270234B (en) * | 2020-10-20 | 2022-04-19 | 天津大学 | A method for target recognition of transmission line insulators based on aerial images |
-
2021
- 2021-04-27 CN CN202110461261.3A patent/CN113239752B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101567046A (en) * | 2009-06-11 | 2009-10-28 | 北京航空航天大学 | Target recognition method of unmanned aerial vehicle based on minimum circle-cover matching |
| CN108761237A (en) * | 2018-05-29 | 2018-11-06 | 福州大学 | Unmanned plane electric inspection process image vital electrical component diagnoses automatically and labeling system |
Non-Patent Citations (1)
| Title |
|---|
| "赋权图的最小环路遍历路径分析与研究";索红军;《渭南师范学院学报》;第27卷(第10期);第78-80页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113239752A (en) | 2021-08-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109685066B (en) | Mine target detection and identification method based on deep convolutional neural network | |
| CN107609557B (en) | Pointer instrument reading identification method | |
| CN103324937B (en) | The method and apparatus of label target | |
| EP3806064B1 (en) | Method and apparatus for detecting parking space usage condition, electronic device, and storage medium | |
| CN111461209B (en) | Model training device and method | |
| CN102915433B (en) | Character combination-based license plate positioning and identifying method | |
| CN107688811B (en) | License plate recognition method and device | |
| CN114998815B (en) | Traffic vehicle identification tracking method and system based on video analysis | |
| CN117372956A (en) | Method and device for detecting state of substation screen cabinet equipment | |
| CN113850799A (en) | A YOLOv5-based Workpiece Detection Method for Micro DNA Extraction Workstations | |
| CN113538585B (en) | High-precision multi-target intelligent identification, positioning and tracking method and system based on unmanned aerial vehicle | |
| CN114012722A (en) | Mechanical arm target grabbing method based on deep learning and edge detection | |
| CN116704490B (en) | License plate recognition method, license plate recognition device and computer equipment | |
| CN116704512A (en) | A meter recognition method and system integrating semantic and visual information | |
| CN113239752B (en) | Unmanned aerial vehicle aerial image automatic identification system | |
| CN114694130A (en) | Method and device for detecting telegraph poles and pole numbers along railway based on deep learning | |
| CN110942008B (en) | Deep learning-based face sheet information positioning method and system | |
| CN113191351B (en) | Method and device for number recognition of digital electric meter, model training method and device | |
| CN116309836A (en) | Three-dimensional object pose recognition method and device based on visual image and electronic equipment | |
| CN112633116A (en) | Method for intelligently analyzing PDF (Portable document Format) image-text | |
| CN117372510B (en) | Map annotation identification method, terminal and medium based on computer vision model | |
| CN112132115B (en) | Image screening method and device | |
| CN114140392B (en) | A method, device and system for automatically determining the maintenance quality of packaging equipment based on deep learning | |
| CN113536860B (en) | Key frame extraction method, and vectorization method of road traffic equipment and facilities | |
| CN110738209B (en) | License plate detection method based on deep learning |
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 |