CN106875407B - Unmanned aerial vehicle image canopy segmentation method combining morphology and mark control - Google Patents
Unmanned aerial vehicle image canopy segmentation method combining morphology and mark control Download PDFInfo
- Publication number
- CN106875407B CN106875407B CN201710070334.XA CN201710070334A CN106875407B CN 106875407 B CN106875407 B CN 106875407B CN 201710070334 A CN201710070334 A CN 201710070334A CN 106875407 B CN106875407 B CN 106875407B
- Authority
- CN
- China
- Prior art keywords
- image
- canopy
- remote sensing
- segmentation
- forest
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 47
- 230000011218 segmentation Effects 0.000 title claims abstract description 39
- 238000001914 filtration Methods 0.000 claims abstract description 19
- 230000000877 morphologic effect Effects 0.000 claims abstract description 17
- 238000006243 chemical reaction Methods 0.000 claims abstract description 12
- 238000012545 processing Methods 0.000 claims abstract description 10
- 239000003550 marker Substances 0.000 claims description 15
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000000926 separation method Methods 0.000 claims description 6
- 230000000007 visual effect Effects 0.000 claims description 4
- 238000005260 corrosion Methods 0.000 claims description 3
- 230000007797 corrosion Effects 0.000 claims description 3
- 230000003628 erosive effect Effects 0.000 claims description 3
- 230000005855 radiation Effects 0.000 claims description 3
- 238000007796 conventional method Methods 0.000 abstract description 3
- 230000009466 transformation Effects 0.000 abstract 1
- 238000001514 detection method Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 230000003044 adaptive effect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 244000025254 Cannabis sativa Species 0.000 description 2
- 241000218631 Coniferophyta Species 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 238000011835 investigation Methods 0.000 description 2
- 238000011430 maximum method Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 240000000731 Fagus sylvatica Species 0.000 description 1
- 235000010099 Fagus sylvatica Nutrition 0.000 description 1
- 240000007817 Olea europaea Species 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000009395 breeding Methods 0.000 description 1
- 230000001488 breeding effect Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 239000013589 supplement Substances 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
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20036—Morphological image processing
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30188—Vegetation; Agriculture
Landscapes
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to an unmanned aerial vehicle image canopy segmentation method combining morphology and mark control, which comprises the following steps: acquiring a plurality of local remote sensing images of the forest area by using an unmanned aerial vehicle, and carrying out mosaic and orthorectification to obtain a complete remote sensing image; carrying out smooth filtering processing on a green light wave band by adopting a Gaussian filtering method; detecting the position of a crown top point from a green light wave band by adopting a self-adaptive local maximum value searching method; imposing the acquired canopy vertex position information on the image through a forced minimum value conversion by using morphological operation; for the true-color remote sensing image subjected to orthorectification, obtaining a binary image only containing a canopy area and a non-canopy area by adopting an ISODATA clustering algorithm, and taking the extracted non-canopy area as a segmented external mark; and (4) adding the external mark to the image after the forced minimum value conversion to perform watershed transformation and segmentation to obtain accurate forest stand single tree crown boundary information. The method effectively solves the problem of inaccurate forest canopy boundary segmentation caused by the conventional method.
Description
Technical Field
The invention relates to an unmanned aerial vehicle image canopy segmentation method combining morphology and mark control.
Background
The tree crown is used as a place for obtaining light energy and converting the energy, and has important significance for researching forest growth conditions and dynamic changes. But due to the complexity and randomness of the forest structure, the acquisition of the tree crown shape and the boundary information is extremely difficult. In recent years, with the development of satellite remote sensing technology, especially the gradual improvement of the spatial resolution of satellite images, the estimation precision of forest crowns is improved, but the obtained remote sensing data is far from meeting the requirements of forestry investigation due to the influences of factors such as climate, period, resolution, cost and the like. Unmanned aerial vehicle remote sensing is taken as a new remote sensing technology, has the advantages of special maneuverability, timeliness and low cost, is easy to obtain data and the like, gradually becomes an effective supplement means of the satellite remote sensing technology, and is widely applied to multiple fields. Although research on unmanned aerial vehicle technology is increasing, research on extracting forest canopy structure information from visible light unmanned aerial vehicle images is still in a test stage, for example, the degree of reliability of acquiring crown parameters from common unmanned aerial vehicle camera images is evaluated by the terminal, and a test is performed on an olive breeding park in the spanish koldowa area, and the RMSE of crown width estimation reaches 0.28. Chianucci and the like utilize a true color image acquired by an eBee unmanned flight system, and estimate the forest canopy coverage of the beech forest by combining an LAB2 image classification method, wherein the decision coefficient R2 reaches about 0.7; in addition, Morgenroth, Mathews and the like utilize point cloud data generated by unmanned aerial vehicle images to analyze forest canopy structures, and certain results are obtained. However, the conventional forest crown segmentation method causes the problem of inaccurate forest crown boundary segmentation, which brings uncertainty to the precision of the forest parameter remotely acquired by the unmanned aerial vehicle.
Disclosure of Invention
In view of this, the present invention provides an unmanned aerial vehicle image canopy segmentation method combining morphology and marker control, which effectively solves the problem of inaccurate canopy boundary segmentation caused by the conventional method.
In order to achieve the purpose, the invention adopts the following technical scheme: an unmanned aerial vehicle image canopy segmentation method combining morphology and mark control is characterized by comprising the following steps:
step S1: acquiring a plurality of local remote sensing images of the forest zones by using an unmanned aerial vehicle, and carrying out mosaic and orthorectification on the plurality of remote sensing images of the forest zones to obtain complete remote sensing images of the forest zones;
step S2: carrying out smooth filtering processing on a green light wave band of the complete remote sensing image by adopting a Gaussian filtering method;
step S3: detecting the position of a crown vertex from a green light wave band of the complete remote sensing image by adopting a self-adaptive local maximum value searching method;
step S4: imposing the acquired canopy vertex position information on the green light wave band image after smooth filtering through one forced minimum value conversion by using morphological operation;
step S5: for the complete remote sensing image obtained in the step S1, obtaining a binary image only containing a canopy area and a non-canopy area by adopting an ISODATA clustering algorithm, and taking the extracted non-canopy area as a segmented external mark;
step S6: based on the results obtained in step S4 and step S5, an external marker is imposed on the image after the forced minimum value conversion for watershed transform segmentation, and accurate forest stand individual tree canopy boundary information is obtained.
Furthermore, the local remote sensing image is a true color image, and the resolution is between 0.05 and 0.20 m.
Further, the specific method of step S2 is as follows: a Gaussian distribution curve is adopted to process a green light wave band of the complete remote sensing image, the noise level of the image is reduced, the radiation intensity value of the top of the canopy is strengthened, and a filter formula is as follows:
in the formula, G (i, j) is the Gaussian filter result of the image pixel at the ith row and the j column, i and j are natural numbers, sigma is the mean square error of the Gaussian filter, and the minimum canopy size in the forest stand is selected as a window for image filtering processing by sigma value taking.
Further, the specific method of step S3 is as follows:
step S31: detecting the position of a potential canopy peak in the sample plot through a fixed window to obtain the potential canopy peak;
step S32: judging the obtained potential crown vertex by adopting a self-adaptive dynamic window, if the current vertex is the maximum value of the corresponding window area, storing, and if not, deleting; the size of the dynamic window is determined by calculating the variation range value of the half variance of the eight section directions of the potential vertexes, and the half variance calculation formula of the image pixel is as follows:
wherein γ (h) is the empirical half-variance value, xiIs the pixel position of the image, h is the spatial separation distance of two pixels, Z (x) is the corresponding image xiWhere the pixel value, N is the logarithm of the pixel pair at a certain separation distance.
Further, the specific method of step S4 is as follows:
step S41: performing negation operation on the filtered image f, and then performing minimum value marking on the acquired crown vertex to obtain a marked image, wherein the formula is as follows:
in the formula (f)mFor the acquired marked image, p is each pixel of the image, tmaxIs the maximum value of the image;
step S42: pixel-by-pixel calculation image f +1 and mark image fmThe minimum value between the two is used for carrying out forced minimum value conversion on the image;
in this step, the morphological calculation is represented as: (f +1) ^ fmThen using the marker image fmFor (f +1) ^ fmPerforming morphological erosion reconstruction as follows:
in the formula (f)mpIn order to force the minimum transformed image,
description of (f +1) ^ fmIn the mark image fmAnd (5) reconstructing the morphological corrosion under the mask.
Further, the specific method of step S5 is as follows: classifying by adopting an ISODATA (inter-digital image data) unsupervised classification method based on the obtained complete remote sensing image, wherein the set classification number is more than or equal to 10, and the maximum iteration number is more than or equal to 5; and merging the obtained classification results through visual interpretation to obtain a binary image only containing the two types of the crown areas and the non-crown areas, and taking the extracted non-crown areas as external marks for segmentation.
Further, the watershed transform segmentation in step S6 adopts the following formula:
where WaterShored () is a WaterShed function; mask is a Mask function; b isOutMaskIs an external mark, i.e. a non-canopy area, WcanopyIs the boundary of the single-tree canopy of the forest stand.
Compared with the prior art, the invention has the following beneficial effects: the method effectively solves the problem of inaccurate forest canopy boundary segmentation caused by the conventional method; the method is beneficial to the quick and effective extraction of the forest crown information, and provides powerful support for the accurate and efficient estimation of the forest branch number and the canopy density in the forest resource investigation.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
Fig. 2A is an image of an unmanned aerial vehicle according to a first embodiment of the present invention.
Fig. 2B shows the green band filtering result according to the first embodiment of the present invention.
FIG. 2C is a direct watershed segmentation result according to a first embodiment of the present invention.
Fig. 2D shows a crown vertex extracted by using a fixed window according to an embodiment of the present invention.
FIG. 2E shows the result of removing outliers using a variable window according to an embodiment of the present invention.
Fig. 2F is a non-canopy binary image of a canopy according to the first embodiment of the present invention.
Fig. 2G shows the morphological reconstruction labeling result according to the first embodiment of the present invention.
Fig. 2H shows the result of adding the internal and external marks according to the first embodiment of the present invention.
Fig. 2I shows the result of image segmentation with internal and external labels according to a first embodiment of the present invention.
Fig. 3A illustrates an unmanned aerial vehicle image according to a second embodiment of the present invention.
FIG. 3B shows the direct watershed segmentation result of the second embodiment of the present invention.
Fig. 3C illustrates canopy anchor points extracted using a fixed window according to a second embodiment of the present invention.
FIG. 3D is a diagram illustrating an adaptive window exception removal result according to a second embodiment of the present invention.
Fig. 3E illustrates an image of a vertex of a forced canopy according to a second embodiment of the present invention.
FIG. 3F shows the morphological reconstruction results of example two of the present invention.
Fig. 3G is a non-canopy binary image of a canopy according to a second embodiment of the present invention.
FIG. 3H shows the direct watershed segmentation result of the intra-labeled image according to the second embodiment of the present invention.
FIG. 3I shows the result of image segmentation with internal and external labels according to the second embodiment of the present invention.
Detailed Description
The invention is further explained below with reference to the drawings and the embodiments.
Referring to fig. 1, the invention provides a method for segmenting a canopy of an unmanned aerial vehicle image by combining morphology and marker control, which is characterized by comprising the following steps:
step S1: acquiring a plurality of local remote sensing images of the forest area with the resolution ratio of 0.05-0.20m by using an unmanned aerial vehicle, and carrying out mosaic and orthorectification on the plurality of forest area remote sensing images to obtain a complete remote sensing image of the forest area; the local remote sensing image is a true color image at least comprising red, green and blue wave bands, the course and the lateral overlapping rate of the image are more than or equal to 80%, and the complete remote sensing image obtained by inlaying and orthoscopic correction has no obvious splicing trace.
Step S2: carrying out smooth filtering processing on a green light wave band of the complete remote sensing image by adopting a Gaussian filtering method; the specific method comprises the following steps: a Gaussian distribution curve (bell-shaped curve) is adopted to process a green light wave band of the complete remote sensing image, the noise level of the image is reduced, and the radiation intensity value of the top point of the canopy is strengthened, wherein the filter formula is as follows:
in the formula, G (i, j) is the Gaussian filter result of the image pixel at the ith row and the j column, i and j are natural numbers, sigma is the mean square error of the Gaussian filter, and the minimum canopy size in the forest stand is selected as a window for image filtering processing by taking the sigma value.
Step S3: detecting the position of a crown vertex from a green light wave band of the complete remote sensing image by adopting a self-adaptive local maximum value searching method; the specific method comprises the following steps:
step S31: firstly, detecting the position of a potential canopy peak in a sample plot through a small fixed window to obtain the potential canopy peak;
step S32: judging the obtained potential crown vertex by adopting a self-adaptive dynamic window, if the current vertex is the maximum value of the corresponding window area, storing, and if not, deleting; the size of the dynamic window is determined by calculating the variation range value of the half variance of the eight section directions of the potential vertexes, and the half variance calculation formula of the image pixel is as follows:
wherein γ (h) is the empirical half-variance value, xiIs the pixel position of the image, h is the spatial separation distance of two pixels, Z (x) is the corresponding image xiWhere the pixel value, N is the logarithm of the pixel pair at a certain separation distance.
Step S4: imposing the acquired canopy vertex position information on the green light wave band image after smooth filtering through one forced minimum value conversion by using morphological operation; the specific method comprises the following steps:
step S41: firstly, performing negation operation on an image f after smoothing filtering processing, and then performing minimum value marking on an acquired crown vertex to obtain a marked image, wherein the formula is as follows:
in the formula (f)mFor the acquired marked image, p is each pixel of the image, tmaxIs the maximum value of the image;
step S42: then, the pixel-by-pixel calculation image f +1 and the marker image fmThe minimum value between the two is used for carrying out forced minimum value conversion on the image;
in this step, the morphological calculation is represented as: (f +1) ^ fmThen using the marker image fmFor (f +1) ^ fmPerforming morphological erosion reconstruction as follows:
in the formula (f)mpIn order to force the minimum transformed image,
description of (f +1) ^ fmIn the mark image fmAnd (5) reconstructing the morphological corrosion under the mask.
Step S5: for the complete remote sensing image obtained in the step S1, obtaining a binary image only containing a canopy area and a non-canopy area by adopting an ISODATA clustering algorithm, and taking the extracted non-canopy area as a segmented external mark; the specific method comprises the following steps: classifying by adopting an ISODATA (inter-digital image data) unsupervised classification method based on the obtained complete remote sensing image, wherein the set classification number is more than or equal to 10, and the maximum iteration number is more than or equal to 5; and merging the obtained classification results through visual interpretation to obtain a binary image only containing the two types of the crown areas and the non-crown areas, and taking the extracted non-crown areas as external marks for segmentation.
Step S6: based on the results obtained in step S4 and step S5, an external marker is imposed on the image after the forced minimum value conversion for watershed transform segmentation, and accurate forest stand individual tree canopy boundary information is obtained. The watershed transform segmentation adopts the following formula:
where WaterShored () is a WaterShed function; mask is a Mask function; b isOutMaskIs an external mark, i.e. a non-canopy area, WcanopyIs the boundary of the single-tree canopy of the forest stand.
In order to make the technical solution of the present invention better understood, the present invention is described in detail below with reference to two embodiments. The local remote sensing image acquired by the unmanned aerial vehicle is an RGB true-color image, preprocessing is performed by PIX4D software, and after mosaic and orthorectification, the image resolution is 7 cm.
The first embodiment is as follows:
FIG. 2A shows the original visible image of plot 1, plot 1 is a conifer plot, and there are isolated crowns and overlapping crowns. Fig. 2B shows the results of maximum filtering and gaussian smoothing filtering of the green band, which enhances the spectral difference between canopy and non-canopy and reduces the spectral heterogeneity inside canopy. Fig. 2C shows the phenomenon of over-segmentation when the green band after the filtering process is directly subjected to watershed segmentation. This is because there are some noise values in the image except for the crown apex, and there are roads and open spaces in the image;
FIG. 2D shows a crown vertex detection result by applying a fixed window local maximum method, where there is a problem that a plurality of vertices are detected by a part of crowns;
FIG. 2E shows the crown vertex detection result by applying the maximum variable window (adaptive window) method based on the fixed window detection result, which can be found to eliminate the phenomenon that a plurality of vertices appear in part of the crown;
FIG. 2F is a diagram of a forest crown and non-forest crown binary image obtained by unsupervised classification;
FIG. 2G: the method is characterized in that a green light wave band is subjected to morphological reconstruction and forced minimum conversion processing, and the found crown vertex marks are forced on an image, so that watershed segmentation is ensured to be performed only according to the tree vertex marks;
FIG. 2H is the result of adding an external marker of a non-canopy region to the result of FIG. 2G;
FIG. 2I shows the result of watershed segmentation using the inside and outside labels of the crown, where each closed polygon represents a crown. By overlapping the crown delineation result with the original image, the result obtained by the algorithm is relatively good.
Example two:
fig. 3A is an original visible image of plot 2, plot 2 being a broadleaf forest plot. FIG. 3B shows the direct watershed segmentation of the green band. An over-segmentation phenomenon occurs. This is because there are some noise values in the image besides the crown vertices, and the grass fields in the image;
FIG. 3C shows the crown vertex detection result by applying the fixed window local maximum method, which has the problem that part of the crowns detect multiple vertices and the problem that the vertices are detected on the grassland;
FIG. 3D shows the top point of the canopy detected by applying the maximum value method of the variable window (adaptive window) based on the fixed window detection result, which can find that the top point problem of the grassland still exists, but the phenomenon that a plurality of top points appear on part of the canopy is eliminated;
FIG. 3E shows the result of the resulting crown vertex mark being imposed on the green band image; FIG. 3F is a morphological reconstruction result that ensures that watershed segmentations will only be segmented according to the tree-top labels. In addition, in order to eliminate the influence of the non-canopy region on the canopy segmentation boundary, the segmentation result needs to be subjected to mask processing. FIG. 3G is a non-canopy binary image of canopy obtained by unsupervised classification. FIG. 3H shows the result of directly performing watershed segmentation on the crown-labeled image, which fails to eliminate the influence of shrub and grass land; FIG. 3I is a diagram of a result of watershed segmentation based on a tree top marker and a non-canopy mask marker, which is relatively good.
Based on the above experimental segmentation results, the results of the individual forest crown segmentation and the visual interpretation of the two samples were compared and verified, and the results are shown in table 1.
TABLE 1 sample area single tree crown extraction accuracy analysis
From Table 1, it can be seen that the crown division accuracy of both plots is high, and plot-01 of conifer forest is 94.54%, while plot-02 of broadleaf forest is 95.56%.
The above description is only a preferred embodiment of the present invention, and all equivalent changes and modifications made in accordance with the claims of the present invention should be covered by the present invention.
Claims (4)
1. An unmanned aerial vehicle image canopy segmentation method combining morphology and mark control is characterized by comprising the following steps:
step S1: acquiring a plurality of local remote sensing images of the forest area with the resolution ratio of 0.05-0.20m by using an unmanned aerial vehicle, and carrying out mosaic and orthorectification on the plurality of forest area remote sensing images to obtain a complete remote sensing image of the forest area; the local remote sensing image is a true color image at least comprising red, green and blue wave bands, and the course and the lateral overlapping rate of the image are more than or equal to 80 percent;
step S2: carrying out smooth filtering processing on a green light wave band of the complete remote sensing image by adopting a Gaussian filtering method;
step S3: detecting the position of a crown vertex from a green light wave band of the complete remote sensing image by adopting a self-adaptive local maximum value searching method;
step S4: imposing the acquired canopy vertex position information on the green light wave band image after smooth filtering through one forced minimum value conversion by using morphological operation;
step S5: for the complete remote sensing image obtained in the step S1, obtaining a binary image only containing a canopy area and a non-canopy area by adopting an ISODATA clustering algorithm, and taking the extracted non-canopy area as a segmented external mark;
step S6: based on the results obtained in the steps S4 and S5, the external marker is added to the image after the forced minimum value conversion for watershed transform segmentation, so as to obtain accurate forest stand single tree canopy boundary information;
the specific method of step S3 is as follows:
step S31: detecting the position of a potential canopy peak in the sample plot through a fixed window to obtain the potential canopy peak;
step S32: judging the obtained potential crown vertex by adopting a self-adaptive dynamic window, if the current vertex is the maximum value of the corresponding window area, storing, and if not, deleting; the size of the dynamic window is determined by calculating the variation range value of the half variance of the eight section directions of the potential canopy vertexes, and the half variance calculation formula of the image pixel is as follows:
wherein γ (h) is the empirical half-variance value, xiIs the pixel position of the image, h is the spatial separation distance of two pixels, Z (x)i) Is corresponding to the image xiThe pixel value, N, is the logarithm of the pixel pair at a certain separation distance;
the specific method of step S4 is as follows:
step S41: performing negation operation on the filtered image f, and then performing minimum value marking on the acquired crown vertex to obtain a marked image, wherein the formula is as follows:
in the formula (f)mFor the acquired marked image, p is each pixel of the image, tmaxIs the maximum value of the image;
step S42: pixel-by-pixel calculation image f +1 and mark image fmThe minimum value between the two is used for carrying out forced minimum value conversion on the image;
in this step, the morphological calculation is represented as: (f +1) ^ fmThen using the marker image fmFor (f +1) ^ fmPerforming morphological erosion reconstruction as follows:
in the formula (f)mpIn order to force the minimum transformed image,
description of (f +1) ^ fmIn the mark image fmAnd (5) reconstructing the morphological corrosion under the mask.
2. The unmanned aerial vehicle image canopy segmentation method combining morphology and marker control according to claim 1, wherein: the specific method of step S2 is as follows: a Gaussian distribution curve is adopted to process a green light wave band of the complete remote sensing image, the noise level of the image is reduced, the radiation intensity value of the top of the canopy is strengthened, and a filter formula is as follows:
in the formula, G (i, j) is a Gaussian filter result of image pixels at the ith row and the jth column, i and j are natural numbers, sigma is the mean square error of the Gaussian filter, and the minimum canopy size in a forest stand is selected as a window for image filtering processing by sigma value taking.
3. The unmanned aerial vehicle image canopy segmentation method combining morphology and marker control according to claim 1, wherein: the specific method of step S5 is as follows: classifying by adopting an ISODATA (inter-digital image data) unsupervised classification method based on the obtained complete remote sensing image, wherein the set classification number is more than or equal to 10, and the maximum iteration number is more than or equal to 5; and merging the obtained classification results through visual interpretation to obtain a binary image only containing the two types of the crown areas and the non-crown areas, and taking the extracted non-crown areas as external marks for segmentation.
4. The unmanned aerial vehicle image canopy segmentation method combining morphology and marker control according to claim 1, wherein: in step S6, the watershed transform segmentation is expressed by the following formula:
where WaterShored () is a WaterShed function; mask is a Mask function; b isOutMaskIs an external mark, i.e. a non-canopy area, WcanopyIs the boundary of the single-tree canopy of the forest stand.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710070334.XA CN106875407B (en) | 2017-02-09 | 2017-02-09 | Unmanned aerial vehicle image canopy segmentation method combining morphology and mark control |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710070334.XA CN106875407B (en) | 2017-02-09 | 2017-02-09 | Unmanned aerial vehicle image canopy segmentation method combining morphology and mark control |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106875407A CN106875407A (en) | 2017-06-20 |
| CN106875407B true CN106875407B (en) | 2020-06-12 |
Family
ID=59167303
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710070334.XA Expired - Fee Related CN106875407B (en) | 2017-02-09 | 2017-02-09 | Unmanned aerial vehicle image canopy segmentation method combining morphology and mark control |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106875407B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107516068A (en) * | 2017-07-26 | 2017-12-26 | 福州大学 | A kind of method that single ebon hat is extracted in the high resolution image from unmanned plane |
| CN109446983A (en) * | 2018-10-26 | 2019-03-08 | 福州大学 | A kind of coniferous forest felling accumulation evaluation method based on two phase unmanned plane images |
| CN110189328B (en) * | 2019-06-11 | 2021-02-23 | 北华航天工业学院 | Satellite remote sensing image processing system and processing method thereof |
| CN111160236A (en) * | 2019-12-27 | 2020-05-15 | 北京林业大学 | Automatic dividing method for watershed canopy by combining forest region three-dimensional morphological filtering |
| TWI749833B (en) * | 2020-10-29 | 2021-12-11 | 中華學校財團法人中華科技大學 | UAV sloping soil-rock watershed image identification method, system and application |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102096818A (en) * | 2011-01-31 | 2011-06-15 | 中国林业科学研究院资源信息研究所 | Object-oriented automatic extracting method and system for outline and parameter of remote sensing image crown |
| CN103426200A (en) * | 2013-08-26 | 2013-12-04 | 天津大学 | Tree three-dimensional reconstruction method based on unmanned aerial vehicle aerial photo sequence image |
-
2017
- 2017-02-09 CN CN201710070334.XA patent/CN106875407B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102096818A (en) * | 2011-01-31 | 2011-06-15 | 中国林业科学研究院资源信息研究所 | Object-oriented automatic extracting method and system for outline and parameter of remote sensing image crown |
| CN103426200A (en) * | 2013-08-26 | 2013-12-04 | 天津大学 | Tree three-dimensional reconstruction method based on unmanned aerial vehicle aerial photo sequence image |
Non-Patent Citations (4)
| Title |
|---|
| 基于标记控制分水岭分割方法的高分辨率遥感影像单木树冠提取;郭昱杉 等;《地球信息科学学报》;20160930;第18卷(第9期);第1259-1266页 * |
| 基于标记控制分水岭算法的树冠高程模型分割;刘峰 等;《中国农学通报》;20111231;第29卷(第19期);第126-133页 * |
| 面向对象的退耕还林造林效果遥感特征提取技术研究;陈永富 等;《林业资源管理》;20060430(第2期);第49-54页 * |
| 高空间分辨率遥感的单木树冠自动提取方法与应用;刘晓双 等;《浙江林学院学报》;20100228;第27卷(第1期);第57-61页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106875407A (en) | 2017-06-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106875407B (en) | Unmanned aerial vehicle image canopy segmentation method combining morphology and mark control | |
| Huang et al. | Individual tree crown detection and delineation from very-high-resolution UAV images based on bias field and marker-controlled watershed segmentation algorithms | |
| Zhang et al. | Object-oriented shadow detection and removal from urban high-resolution remote sensing images | |
| CN109558806B (en) | Method for detecting high-resolution remote sensing image change | |
| CN111797712B (en) | Remote sensing image cloud and cloud shadow detection method based on multi-scale feature fusion network | |
| CN108710840B (en) | Visual navigation path identification method for farmland pesticide spraying robot | |
| CN110120046B (en) | Potential landslide identification method integrating DEM (digital elevation model), optical remote sensing and deformation information | |
| WO2018076138A1 (en) | Target detection method and apparatus based on large-scale high-resolution hyper-spectral image | |
| CN111027497B (en) | Weak and small target rapid detection method based on high-resolution optical remote sensing image | |
| CN108537169B (en) | High-resolution remote sensing image road extraction method based on center line and road width detection algorithm | |
| CN110175556B (en) | Remote sensing image cloud detection method based on Sobel operator | |
| CN107992856B (en) | High-resolution remote sensing building shadow detection method under urban scene | |
| CN105139396B (en) | Full-automatic remote sensing image cloud and fog detection method | |
| CN113255452A (en) | Extraction method and extraction system of target water body | |
| CN115731257A (en) | Leaf form information extraction method based on image | |
| CN112115778B (en) | Intelligent lane line identification method under ring simulation condition | |
| Mustafa et al. | RETRACTED: Water surface area detection using remote sensing temporal data processed using MATLAB | |
| CN109785318B (en) | Remote sensing image change detection method based on facial line primitive association constraint | |
| CN113421301B (en) | Method and system for positioning central area of field crop | |
| CN115266719A (en) | Water quality monitoring method and device based on artificial intelligence | |
| Abraham et al. | Unsupervised building extraction from high resolution satellite images irrespective of rooftop structures | |
| CN113989253A (en) | Farmland target object information acquisition method and device | |
| CN110348443B (en) | Multi-feature separation statistical method for tree trunk in complex environment and trunk identification method | |
| Maarir et al. | Roads detection from satellite images based on active contour model and distance transform | |
| Abraham et al. | A fuzzy based automatic bridge detection technique for satellite 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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200612 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |