CN109919088A - A kind of Karst region dragon fruit single plant identification extraction method - Google Patents
A kind of Karst region dragon fruit single plant identification extraction method Download PDFInfo
- Publication number
- CN109919088A CN109919088A CN201910168451.9A CN201910168451A CN109919088A CN 109919088 A CN109919088 A CN 109919088A CN 201910168451 A CN201910168451 A CN 201910168451A CN 109919088 A CN109919088 A CN 109919088A
- Authority
- CN
- China
- Prior art keywords
- dragon fruit
- image
- single plant
- area
- plant
- 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.)
- Granted
Links
Landscapes
- Image Processing (AREA)
Abstract
The invention discloses a kind of Karst region dragon fruit single plants to identify extraction method, method includes the following steps: (1), which takes pictures, obtains dragon fruit aerophotograph image；(2) image procossing, which obtains, surveys area's orthophotoquad；(3) pre-process to image: (4) carry out wave band calculating: (4) are split processing to image and obtain threshold value；(5) the vector patch for surveying area's dragon fruit is obtained by Threshold segmentation Target scalar and background value；(6) judge whether figure spot separates；(7) statistics separates thorough single plant number and calculates the separation number of disjunctor plant；(8) will separate separation number of the thorough plant number plus disjunctor plant is the entire single plant total quantity for surveying area.Karst region single plant dragon fruit image counting accuracy can be greatly improved compared to the mode of conventional satellite remote sensing using dragon fruit single plant extracting method of the invention, realize that dragon fruit single plant automatically extracts batch processing, treatment effeciency is high, reduces labor intensity significantly.
Description
Technical field
The invention belongs to dragon fruit single plant identification technology fields, and it is automatic to be related to a kind of Karst region dragon fruit single plant identification
Extracting method.
Background technique
In order to alleviate that karst plateau Canyon Area agricultural resource is nervous and environmental pressure it is huge between contradiction, grasp exploitation
The utilization of the theory that typical industrial crops are infinitely worth in the confined space, the future of agriculture will make the transition from " extensive " to " fine ".
But the New Remote Sensing Technology application of existing support precision agriculture mainly also focuses in the identification classification of atural object.Therefore this research and development
In conjunction with Guizhou extreme terrain landforms select unmanned aerial vehicle platform carry visible light lens obtain high-resolution image, using contain only R, G,
The high-resolution image of tri- wave band of B carries out wave band calculating and is partitioned into Target scalar, proposes a kind of with plant average area segmentation strain clump
Thought combination visual programming spatial modeling tool model composer, built dragon fruit single plant and automatically extracted batch processing and essence
Degree verifying model.
Summary of the invention
The technical problem to be solved by the present invention is providing a kind of Karst region dragon fruit single plant identification side of automatically extracting
Method, to solve the technical problems existing in the prior art.
The technical scheme adopted by the invention is as follows: a kind of Karst region dragon fruit single plant identification extraction method, the party
Method the following steps are included:
(1) visible light lens are carried using unmanned aerial vehicle platform and plans automatic course line, by unmanned plane flying height, course, other to weight
Folded degree is respectively set to 40m, 70%, 60%, take pictures obtaining aerophotograph；
(2) aerophotograph is imported using image joint software, completes aerial triangulation calculating generation boat and take the photograph point cloud data, obtains
The area Qu Ce digital orthophoto map；
(3) pass through the deformation, distortion, distortion, fuzzy and noise to generating in aerophotograph acquisition process by unmanned plane shake
It corrects, and carries out pretreatment acquisition survey area's high resolution image of image enhancement, ornamenting, cutting, reconstruction to it；
(4) wave band calculating, calculation method are as follows: float (2*b2-b1-b3/2*b2+ are carried out to the high resolution image of acquisition
B1+b3), wherein float indicates that calculated result is floating type, and b1, b2, b3 respectively indicate three wave band of red, green, blue；
(5) background and target two parts are classified as according to the gamma characteristic of image using maximum kind variance method, utilize IDL
Interactive language programming, which is realized, automatically extracts segmentation threshold, when threshold value T makes the class variance maximum between target and background, threshold value
T is the optimal threshold of segmentation object atural object；
(6) it is divided into two different figure layers to obtain from background value Target scalar using the threshold value of extraction and surveys area's dragon fruit
Vector patch；
(7) area that each fritter figure spot is calculated using the geometry tool of ArcGIS software, rejects background and scrappy figure spot
Obtain the vector patch for containing only Target scalar；
(8) judge whether figure spot separates, counted to thorough patch is separated as dragon fruit single plant number, not to separation
The strain clump thoroughly to connect together, a kind of method that thorough dragon fruit single plant average area segmentation disjunctor strain clump is separated using use,
Method are as follows: with the area of each disjunctor patch respectively divided by the average area of single plant dragon fruit, obtain point of each disjunctor patch
From number, the single plant that integer is disjunctor plant is converted using the rule of setting by it and separates quantity；
(9) will separate separation number of the thorough plant number plus disjunctor plant is the entire single plant total quantity for surveying area, is obtained
The statistics sum of the area get Ce Target scalar single plant separation；
(10) verifying of human-computer interaction field, which obtains, surveys the precision that the practical single plant number verifying in area automatically extracts, and verification method is
In formula, ρ indicates accuracy, and M indicates to extract strain number, and N indicates practical strain number, has calculated identification by the formula and has mentioned
The dragon fruit precision taken, convenient for evaluating extraction result.
Maximum kind variance method calculation method in step (5):
N0+N1=MN (9)
W0+W1=1 (10)
μ=W0μ0+W1μ1 (11)
σ=W0(μ0-μ)2+W1(μ1-μ)2 (12)
In formula: W0The ratio of whole scape image, W are accounted for for goal pels point1The ratio of whole scape image, μ are accounted for for backdrop pels point0
For Target scalar pixel average gray, μ1For backdrop pels average gray, μ is image overall average gray scale, and MN indicates image size,
N0It is less than the pixel number of T, N for gray scale1It is greater than the pixel number of T for gray scale, σ is inter-class variance.
Beneficial effects of the present invention: compared with prior art, effect of the invention is as follows:
1) it can be greatly improved using dragon fruit single plant extracting method of the invention compared to the mode of conventional satellite remote sensing
Karst region single plant dragon fruit image counting accuracy realizes that dragon fruit single plant automatically extracts batch processing, and treatment effeciency is high, greatly
It reduces labor intensity greatly；
2) high definition boat is obtained by the way that unmanned plane flying height, course, sidelapping degree are respectively set to 40m, 70%, 60%
Piece makes up the cloudy rainy day gas conventional satellite in Karst region (such as Guizhou) and obtains the inadequate defect of image precision, accomplishes to obtain on demand
Take high-resolution image；
3) it is calculated by wave band, obtains the grayscale image of image, with the vegetation index for containing only red, Lu, blue three wave bands building
Operation, compared to multispectral, EO-1 hyperion cost is lower；
4) divide disjunctor plant with this method, disjunctor plant has more fully been divided into single plant.Advantage: for the first time
It proposes the thought for going segmentation strain clump with plant average area, and tests and achieve certain effect；
5) the extraction result that method of the invention obtains can be dragon fruit Growing state survey, dragon fruit the yield by estimation, dragon fruit plant
Morphologic information acquisition, further exploitation dragon fruit is worth and is the service of Karst Mountain Areas precision agriculture, provides certain ginseng
Examine value.
Detailed description of the invention
Fig. 1 is technology path flow chart；
Fig. 2 automatically extracts for dragon fruit single plant and precision test illustraton of model；
Fig. 3 is trial zone unmanned plane visible image figure；
Fig. 4 is wave band calculated result figure；
Fig. 5 is dragon fruit extraction effect figure.
Specific embodiment
With reference to the accompanying drawing and the present invention is described further in specific embodiment.
Embodiment 1: as Figure 1-Figure 5, a kind of Karst region dragon fruit single plant identification extraction method, this method
The following steps are included:
(1) visible light lens are carried using unmanned aerial vehicle platform and plans automatic course line, flying height, course, sidelapping degree are set
It is respectively set to 40m, 70%, 60% and obtains aerophotograph；
(2) aerophotograph is imported using image joint software, completes aerial triangulation calculating generation boat and take the photograph point cloud data, obtains
The area Qu Ce digital orthophoto map；
(3) pass through the deformation, distortion, distortion, fuzzy and noise to generating in aerophotograph acquisition process by unmanned plane shake
It corrects, and carries out pretreatment acquisition survey area's high resolution image of image enhancement, ornamenting, cutting, reconstruction to it；
(4) wave band calculating carried out to the high-resolution image of acquisition, the remote sensing image processing software ENVI of choice of software mainstream,
Tools selection Band Math, calculation method are as follows: float (2*b2-b1-b3/2*b2+b1+b3), wherein float indicates to calculate knot
Fruit is floating type, and b1, b2, b3 respectively indicate three wave band of red, green, blue；
(5) background and target two parts are classified as according to the gamma characteristic of image according to maximum kind variance method (OTSU),
It is realized using the programming of IDL (Interface description language) interactive language and automatically extracts segmentation threshold,
Core concept is when threshold value T makes the class variance maximum between target and background, and threshold value T is the best threshold of segmentation object atural object
Value, the segmentation threshold for obtaining the image is 0.037, and it is high to calculate quickly and precisely degree；
Maximum kind variance method calculation method:
N0+N1=MN (9)
W0+W1=1 (10)
μ=W0μ0+W1μ1 (11)
σ=W0(μ0-μ)2+W1(μ1-μ)2 (12)
In formula: W0The ratio of whole scape image, W are accounted for for goal pels point1The ratio of whole scape image, μ are accounted for for backdrop pels point0
For Target scalar pixel average gray, μ1For backdrop pels average gray, μ is image overall average gray scale, and MN indicates image size,
N0It is less than the pixel number of T, N for gray scale1It is greater than the pixel number of T for gray scale, σ is inter-class variance；
(6) it is divided into two different figure layers to obtain from background value Target scalar using the threshold value of extraction and surveys area's dragon fruit
Vector patch, convenient for quickly recognizing Target scalar；
(7) area that each fritter figure spot is calculated using the geometry tool of ArcGIS software, rejects background and scrappy figure spot
The vector patch for containing only Target scalar is obtained, the VectorLayer for forming complete pure target area is obtained；
(8) judge whether figure spot separates: being counted to thorough patch is separated as dragon fruit single plant number, not to separation
The strain clump thoroughly to connect together proposes a kind of thought for separating thorough dragon fruit single plant average area segmentation disjunctor strain clump, tool
Body is the area with each disjunctor patch respectively divided by the average area of single plant dragon fruit, obtains the separation of each disjunctor patch
Number, this number are double, it is used the rule (regular in the way of rounding up) of setting be converted into integer is to connect
The single plant of body plant separates quantity；
(9) will separate separation number of the thorough plant number plus disjunctor plant is the entire single plant total quantity for surveying area；
(10) verifying of human-computer interaction field, which obtains, surveys the precision that the practical single plant number verifying in area automatically extracts, and verification method is
In formula, ρ indicates accuracy, and M indicates to extract strain number, and N indicates practical strain number.Wherein (7)~(12) are integrated in mould
Type composer realizes automatic batch processing；
(11) outcome evaluation is extracted:
Table 1 extracts each index feature Data-Statistics
Such as table 1, the plant sum automatically extracted is 320, and the practical strain number that the verifying of human-computer interaction field obtains is 295
, the strain number automatically extracted is 25 more than practical strain number, and substituting into precision test formula and can obtaining extraction accuracy is 91.7%, mistake
Rate is 8.3%, causes to mention wrong the reason of dividing more and is mainly derived from the shade of disjunctor plant and the interference of part weeds, wrong by its face
Product has been classified as Target scalar.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain
Within protection scope of the present invention, therefore, protection scope of the present invention should be based on the protection scope of the described claims lid.
Claims (2)
1. a kind of Karst region dragon fruit single plant identifies extraction method, it is characterised in that: method includes the following steps:
(1) visible light lens are carried using unmanned aerial vehicle platform and plans automatic course line, by unmanned plane flying height, course, sidelapping degree
It is respectively set to 40m, 70%, 60%, take pictures obtains the aerophotograph that plantation dragon fruit surveys area；
(2) aerophotograph is imported using image joint software, completes aerial triangulation calculating generation boat and take the photograph point cloud data, obtains and survey
Area's digital orthophoto map；
(3) by aerophotograph acquisition process because of deformation, distortion, distortion, fuzzy and noise correction that unmanned plane shake generates,
And pretreatment acquisition survey area's high resolution image of image enhancement, ornamenting, cutting, reconstruction is carried out to it；
(4) wave band calculating, calculation method are as follows: float (2*b2-b1-b3/2*b2+b1+ are carried out to the high resolution image of acquisition
B3), wherein float indicates that calculated result is floating type, and b1, b2, b3 respectively indicate three wave band of red, green, blue；
(5) background and target two parts are classified as according to the gamma characteristic of image using maximum kind variance method, utilize IDL interaction
Formula Programming with Pascal Language realization automatically extracts segmentation threshold, and when threshold value T makes the class variance maximum between target and background, threshold value T is
For the optimal threshold of segmentation object atural object；
(6) it is divided into two different figure layers to obtain Target scalar and background value using the threshold value of extraction and surveys the arrow of area's dragon fruit
Measure patch；
(7) area that each fritter figure spot is calculated using the geometry tool of ArcGIS software, rejects background and scrappy figure spot obtains
Contain only the vector patch of Target scalar；
(8) judge whether figure spot separates, counted to thorough patch is separated as dragon fruit single plant number, separation is not thorough
The strain clump to connect together, using a kind of with the method for separating thorough dragon fruit single plant average area segmentation disjunctor strain clump, method
Are as follows: with the area of each disjunctor patch respectively divided by the average area of single plant dragon fruit, the separation number of each disjunctor patch is obtained,
The single plant that integer is disjunctor plant, which is converted, using the rule of setting by it separates quantity；
(9) will separate separation number of the thorough plant number plus disjunctor plant is the entire single plant total quantity for surveying area；
(10) verifying of human-computer interaction field, which obtains, surveys the precision that the practical single plant number verifying in area automatically extracts, and verification method is
In formula, ρ indicates accuracy, and M indicates to extract strain number, and N indicates practical strain number.
2. a kind of Karst region dragon fruit single plant according to claim 1 identifies extraction method, it is characterised in that:
Maximum kind variance method calculation method in step (5):
N0+N1=MN (9)
W0+W1=1 (10)
μ=W0μ0+W1μ1 (11)
σ=W0(μ0-μ)2+W1(μ1-μ)2 (12)
In formula: W0The ratio of whole scape image, W are accounted for for goal pels point1The ratio of whole scape image, μ are accounted for for backdrop pels point0For mesh
Mark atural object pixel average gray, μ1For backdrop pels average gray, μ is image overall average gray scale, and MN indicates image size, N0For
Gray scale is less than the pixel number of T, N1It is greater than the pixel number of T for gray scale, σ is inter-class variance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910168451.9A CN109919088B (en) | 2019-03-06 | 2019-03-06 | Automatic extraction method for identifying individual plants of pitaya in karst region |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910168451.9A CN109919088B (en) | 2019-03-06 | 2019-03-06 | Automatic extraction method for identifying individual plants of pitaya in karst region |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109919088A true CN109919088A (en) | 2019-06-21 |
| CN109919088B CN109919088B (en) | 2022-09-23 |
Family
ID=66963498
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910168451.9A Active CN109919088B (en) | 2019-03-06 | 2019-03-06 | Automatic extraction method for identifying individual plants of pitaya in karst region |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109919088B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110598619A (en) * | 2019-09-06 | 2019-12-20 | 中国农业科学院农业资源与农业区划研究所 | Method and system for identifying and counting fruit trees by using unmanned aerial vehicle images |
| CN111798433A (en) * | 2020-07-08 | 2020-10-20 | 贵州师范大学 | Method for identifying and counting mature dragon fruits in plateau mountain area based on unmanned aerial vehicle remote sensing |
| CN113554675A (en) * | 2021-07-19 | 2021-10-26 | 贵州师范大学 | Edible fungus yield estimation method based on unmanned aerial vehicle visible light remote sensing |
| CN114926374A (en) * | 2022-07-21 | 2022-08-19 | 四川新迎顺信息技术股份有限公司 | Image processing method, device and equipment based on AI and readable storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017077543A1 (en) * | 2015-11-08 | 2017-05-11 | Agrowing Ltd | A method for aerial imagery acquisition and analysis |
| CN108022245A (en) * | 2017-12-06 | 2018-05-11 | 南京师范大学 | Photovoltaic panel template automatic generation method based on upper thread primitive correlation model |
-
2019
- 2019-03-06 CN CN201910168451.9A patent/CN109919088B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017077543A1 (en) * | 2015-11-08 | 2017-05-11 | Agrowing Ltd | A method for aerial imagery acquisition and analysis |
| CN108369635A (en) * | 2015-11-08 | 2018-08-03 | 阿格洛英公司 | The method with analysis is obtained for aerial image |
| CN108022245A (en) * | 2017-12-06 | 2018-05-11 | 南京师范大学 | Photovoltaic panel template automatic generation method based on upper thread primitive correlation model |
Non-Patent Citations (1)
| Title |
|---|
| 韩文霆等: "基于无人机遥感技术的玉米种植信息提取方法研究", 《农业机械学报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110598619A (en) * | 2019-09-06 | 2019-12-20 | 中国农业科学院农业资源与农业区划研究所 | Method and system for identifying and counting fruit trees by using unmanned aerial vehicle images |
| CN111798433A (en) * | 2020-07-08 | 2020-10-20 | 贵州师范大学 | Method for identifying and counting mature dragon fruits in plateau mountain area based on unmanned aerial vehicle remote sensing |
| CN111798433B (en) * | 2020-07-08 | 2024-04-26 | 贵州师范大学 | Method for identifying and counting mature dragon fruits in mountain area of plateau based on unmanned aerial vehicle remote sensing |
| CN113554675A (en) * | 2021-07-19 | 2021-10-26 | 贵州师范大学 | Edible fungus yield estimation method based on unmanned aerial vehicle visible light remote sensing |
| CN114926374A (en) * | 2022-07-21 | 2022-08-19 | 四川新迎顺信息技术股份有限公司 | Image processing method, device and equipment based on AI and readable storage medium |
| CN114926374B (en) * | 2022-07-21 | 2022-10-11 | 四川新迎顺信息技术股份有限公司 | Image processing method, device and equipment based on AI and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109919088B (en) | 2022-09-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109919088A (en) | A kind of Karst region dragon fruit single plant identification extraction method | |
| CN110866079A (en) | Intelligent scenic spot real scene semantic map generating and auxiliary positioning method | |
| CN108195767B (en) | Estuary wetland foreign species monitoring method | |
| CN112560623B (en) | Unmanned aerial vehicle-based rapid mangrove plant species identification method | |
| CN112818925B (en) | Urban building and crown identification method | |
| CN103645480A (en) | Geographic and geomorphic characteristic construction method based on laser radar and image data fusion | |
| CN109829425B (en) | Farmland landscape small-scale ground feature classification method and system | |
| CN111814563B (en) | Method and device for classifying planting structures | |
| CN110765977A (en) | Method for extracting wheat lodging information based on multi-temporal remote sensing data of unmanned aerial vehicle | |
| CN109977802A (en) | Crops Classification recognition methods under strong background noise | |
| CN112907520B (en) | Single tree crown detection method based on end-to-end deep learning method | |
| CN108921120A (en) | A kind of cigarette recognition methods adapted under extensive public safety | |
| CN116503705B (en) | Fusion method of digital city multi-source data | |
| CN104751111A (en) | Method and system for recognizing human action in video | |
| CN110889840A (en) | Effectiveness detection method of high-resolution 6 # remote sensing satellite data for ground object target | |
| CN114663814B (en) | Fruit detection and yield estimation method and system based on machine vision | |
| CN110503647A (en) | Wheat plant real-time counting method based on deep learning image segmentation | |
| CN115115954A (en) | Intelligent identification method for pine nematode disease area color-changing standing trees based on unmanned aerial vehicle remote sensing | |
| Zheng et al. | Single shot multibox detector for urban plantation single tree detection and location with high-resolution remote sensing imagery | |
| CN117576559A (en) | Urban greening tree species identification method and system based on orthographic image of unmanned aerial vehicle | |
| US20240290089A1 (en) | Method for extracting forest parameters of wetland with high canopy density based on consumer-grade uav image | |
| CN104008374B (en) | Miner's detection method based on condition random field in a kind of mine image | |
| CN116630828B (en) | Unmanned aerial vehicle remote sensing information acquisition system and method based on terrain environment adaptation | |
| CN102231190A (en) | Automatic extraction method for alluvial-proluvial fan information | |
| Grigillo et al. | Classification based building detection from GeoEye-1 images |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |