CN109977991A - Forest resourceies acquisition method based on high definition satellite remote sensing - Google Patents
Forest resourceies acquisition method based on high definition satellite remote sensing Download PDFInfo
- Publication number
- CN109977991A CN109977991A CN201910064076.3A CN201910064076A CN109977991A CN 109977991 A CN109977991 A CN 109977991A CN 201910064076 A CN201910064076 A CN 201910064076A CN 109977991 A CN109977991 A CN 109977991A
- Authority
- CN
- China
- Prior art keywords
- remote sensing
- image
- forest
- forest land
- information
- 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.)
- Withdrawn
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/001—Texturing; Colouring; Generation of texture or colour
Abstract
Forest resourceies acquisition method based on high definition satellite remote sensing, its method is as follows: the first remotely-sensed data source of selection and the second remotely-sensed data source, it carries out data processing and obtains the first remote sensing image and the second remote sensing image, forest land information extraction and extracting change information are made to the first remote sensing image, for sampling verification, to form forest land investigation raw data, extracting change information is made respectively to the first remote sensing image and the second remote sensing image, and the change information extracted from the two is formed into Forest road hierarchy information, for carrying out database update to database to form Forest road hierarchy monitoring final data, in the present invention, the classification of forest land information and the method extracted are: setting R, G, B respectively indicates the red channel of image, green channel, the average value in three channels of blue channel, utilize R, G, B is as radix, according to different forest land information One-to-one formula is set to classify to forest land information for extracting, the classification of forest land information can be completed rapidly and accurately for extracting.
Description
Technical field
The present invention relates to forest resourceies acquisition method technical fields, provide more particularly to the forestry based on high definition satellite remote sensing
Source acquisition method.
Background technique
The i.e. remote perception of remote sensing (Remote Sensing), refers on certain distance, does not connect directly using detection instrument
Touch target object, recorded from distant place the electromagnetic characteristics of target, by analysis, disclose object characteristic properties and its
The comprehensive Detection Techniques of variation.Photograph is a kind of most common remote sensing, camera not in contact with target subject, but
At intervals, by camera lens the photologging of target subject on egative film, by chemical treatment, photograph just reappears quilt
Take the photograph the image of target.Means used in from photographic subjects to reproducing target are a kind of remote sensing technologies.Remote sensing and other technologies knot
It closes, has the characteristics that in agricultural application scientific, quick, timely.This for make full use of agricultural resource, guiding agricultural production,
Agricultural product equilibrium of supply and demand etc. has great significance.
The application of remote sensing agriculturally mainly carry out agricultural land resource investigation, agricultural output assessment and meteorological disaster,
The monitoring of crop disease and insect, forecast etc..
External Remote Sensing Yield Estimation progress of research situation:
The beginning of Crop Yield Estimation by Remote, United States Department of Agriculture, National Oceanographic and Atmospheric Administration, Space Agency and quotient have been opened first in the U.S.
" Large Area of Crops the yield by estimation experiment (1974~1978) plan " has been formulated in the cooperation of industry portion, has organized and implemented wheat yield estimation plan,
The MSS image handled out using 1~3 reception of Landsat entered the orbit successively is emitted, first to 9 Wheat Productions of big plain in U.S.A
Area, per unit area yield and the yield in state make estimation；Thereafter to including continental United States, Canada and former Soviet Union's some areas wheat flour
Product, per unit area yield and yield make estimation；It is estimated followed by the other regional wheat areas in the world, total output.Investigation and analysis beauty
Wheat cultivation area, emergence situation and the growing way of the main Chan Liang states such as state, the former Soviet Union, Canada, and obtained using meteorological satellite
Meteorological element information, in conjunction with calendar year statistics data carry out comprehensive analysis, the wheat yield estimation model accuracy of foundation be up to 90% with
On.
1980~1986 years, the U.S. formulated " space remote sensing of agricultural and resource is investigated " plan again, and core content is still
It is the cultivated area and the research of yield model of chief crop.Carry out domestic, world's plurality of cereals crop growing state assessment and yield
Forecast.The Chen Shenbin of the comprehensive examination committee of Chinese Academy of Sciences's natural resources is in August, 1992 in foreign country, United States Department of Agriculture agricultural
Office (be responsible for the Crop Estimation of country other than the U.S., and build up operating system) once see the Wheat in China of of that month estimation, corn,
The digital difference -3.53% ,+0.65% and -0.66% that rice total output and later State Statistics Bureau in 1993 announce.This work
Make, obtains huge economic interests in world agricultural trade for the U.S..
Hereafter, also all applied satellite remote sensing technology carries out crops for the states such as the European Community, Russia, France, Japan and India
Growing state survey and yield measuring and calculating, achieve certain achievement.For example, time (since nineteen eighty-three) of the European Community with 10 years,
Build up for agricultural remote sensing application system, nineteen ninety-five in 15 countries, the European Community, 180 scape SPOT images, in conjunction with NOAA shadow
As having carried out agricultural output assessment 60 testing sites, plot and crop species can be accurate to.NASA in 2002 and beauty
The cooperation of the Ministry of Agriculture, state replaces NOAA-AVHRR to carry out Remote Sensing Yield Estimation in Beltsville, Maryland with MODIS data, and MODIS is carried
TERRA satellite be 1999 by the U.S. (National Aeronautics and Space Administration), Japanese (international trade and bureau of industry) and Canada (space
Office, University of Toronto) cooperate transmitting, MODIS data be related to wavelength band wide (36 wave bands), resolution ratio (250,
500,1 000m) it has large improvement than NOAA-AVHRR (5 wave bands, resolution ratio 1100m), these data provide agricultural
Source remote sensing monitoring has higher practical value.
The conversion vegetation for the reflected value that ldso etc. was once obtained with two spectral regions of 500~600nm and 600~700nm refers to
(TV16) is counted to estimate the per unit area yield of wheat and barley, obtaining the related coefficient between yield of wheat and TV16 is 0.78.The same year,
Japanese scientific & technical corporation completes " Remote Sensing Yield Estimation " project, and the precision of Plain agricultural the yield by estimation can be improved, and is conceived to and carries out to the whole world
The yield by estimation.And remote sensing technology has been used for precision agriculture by the U.S., carries out the assessment of region moisture distribution to crops, pest and disease damage is predicted
Deng direct guiding agricultural production.Growing state survey is carried out with satellite remote sensing method and yield estimation has carried out for many years, and method has tended to
It is mature.
Rice yield estimation by remote sensing is in advance and advanced with Asia rice major producing country.China, India, Japan and other countries all into
It went and Remote Sensing Yield Estimation research and obtains preferable effect.Patel and Dash etc. [14] establishes rice yield and the relationship of RVI, examination
It tests area's forecast precision and reaches 96.14%.Miller etc. [15] passes through dry in tiller or fringe stage out with ratio vegetation index
The relationship of substance and per unit area yield estimates per unit area yield.But crop grouting and the stage of ripeness, due between reflectivity and total biomass simultaneously
It is uncorrelated, the canopy biomass of the unpredictable rice of ratio vegetation index.Wiegand, SSRay think by means of normalizing vegetation
Index NDVI { (NIR-R)/(NIR+R) } can forecast production well.
Domestic Remote Sensing Yield Estimation progress situation:
Since " six or five ", China tries out the research that satellite remote sensing carries out crop yield forecast, and carries out in some areas and produce
Measure estimating test.The enforcement period of the seventh five-year plan, National Meteorological Bureau have carried out comprehensive survey of northern 11 provinces and cities' wheat meteorological satellites in 1987 and have produced,
The new method that the Exploration and application period is short, low-cost satellite carries out Crop Estimation.In the project, mainly with long-term meteorology
Based on data, using remote sensing information as measuring means, Remote sensing parameters-crop yield first order recursive mould of different regions is established
Type.1985~1989 years, this project provided the yield forecast of 165 different time and space scales for central and local governments, subtracted for country
Few grain loss adds up economic benefit up to 2,000,000,000 yuan up to 330,000 t or more.
During " eight or five ", Remote Sensing Yield Estimation is classified as brainstorm subject by country, is presided over by the Chinese Academy of Sciences, joint Ministry of Agriculture etc. 40
A unit has been carried out to wheat, corn and rice large area Remote Sensing Yield Estimation experimental study, and large area " Remote Sensing Yield Estimation examination has been built up
Test operating system ", and complete the part basis work of nationwide Remote Sensing Yield Estimation.It was tested by 4 years 1993~1996 years
Operation, respectively to the wheat of four provinces, two city (Hebei, Shandong, Henan, Northern Anhui and Beijing, Tianjin), Hubei, Jiangsu and
The rice in Shanghai City；The monitoring and forecast of the maize sown area, growing way and yield in Jilin Province, in guiding agricultural production and agricultural
Important function has been played in decision.Especially solve key technology problems, it is distant further to carry out nationwide satellite
Sense the yield by estimation provides important guarantee.
Nineteen ninety-five with information system and agriculture feelings speed report, establishes National Resources environment data base；The Chinese Academy of Sciences, weather bureau and
More institution of higher learning, research institute are dedicated to the research of Remote Sensing Yield Estimation technology, and in Zhejiang, Jiangxi, Jiangsu each province and North China, east
North, Jianghan Plain etc. area to the crops such as winter wheat, corn, rice, broom corn millet carry out Remote Sensing Yield Estimation, remote sensing sources choose,
Each sport technique segments such as crop identification, area extraction, model construction, the system integration, which have, significantly to improve.Li Zhe, Zhang Juntao are mentioned
The Maize yeild estimation method combined based on genetic algorithm with artificial neural network out；The it is proposeds such as Hou Yingyu based on crop vegetation
The production estimation model of index and temperature；The Crop Yield Estimation by Remote model based on artificial neural network that east of a river doctor proposes；
Hyperspectral remote sensing estimation models and rice bidirectional reflectance model of the it is proposeds such as royal people's tide professor etc., before these models have been drawn
The selection of the advantages of model, factor of a model are more reasonable, and operability is stronger, and levels of precision is higher.Then, Remote Sensing Yield Estimation method
It has reached its maturity.
The quick obtaining of current land use pattern mainly passes through remote sensing image computer automatic sorting.High-space resolution
Rate image contains richer spatial information compared to low spatial resolution image, and geometrical characteristic and textural characteristics are also brighter
It is aobvious.It therefore, can the easier category attribute information for obtaining atural object using high spatial resolution image.Traditional remote sensing image point
Analysis method is the information extraction technology based on pixel, such as: maximum likelihood method, K- averaging method, iterates self-organizing number at minimum distance method
According to the methods of analysis, these methods are in theory and apply upper all comparative maturities, but these methods only account for the Spectral Properties of image
Reference breath.Blaschke and Strobl thinks that traditional analysis method based on pixel cannot be completely clearly intrinsic with image
Spatial information, be difficult distinguish image present in " the different spectrum of jljl " and " same object different images " phenomenon, in most cases it be
It is a classification by adjacent or neighbouring pixel merger.The shortcomings that in order to overcome traditional technology, needs to research and develop a kind of based on remote sensing
The new forest land information classification approach of image.
Summary of the invention
The forestry based on high definition satellite remote sensing is provided it is an object of the invention to avoid shortcoming in the prior art
Collection of resources method, this method can be completed rapidly and accurately the classification of forest land information for extracting.
The purpose of the present invention is achieved through the following technical solutions:
Forest resourceies acquisition method based on high definition satellite remote sensing is provided, comprising the following steps:
Step 1: the first remotely-sensed data source of selection and the second remotely-sensed data source；
Step 2: the data processing of remote sensing image is carried out to the first remotely-sensed data source and the second remotely-sensed data source, it is distant to obtain first
Feel striograph and the second remote sensing image, the data processing of the remote sensing image includes visual fusion: by the first remotely-sensed data source
Or second the remote sensing image of at least two different times in remotely-sensed data source merged: three wave bands are generated according to multispectral image
Image, Image registration reconnaissance operation before merging, is transformed into HSV image, generates V-band image, and inverse transformation generates RGB image；
Step 3: making forest land information extraction and extracting change information to the first remote sensing image, sampling verification is used for, to form woods
Investigate raw data in ground；
Step 4: extracting change information is made to the first remote sensing image and the second remote sensing image respectively, and will be extracted from the two
Change information out forms Forest road hierarchy information, carries out database update for the database to step 3 to form Forest road hierarchy
Monitor final data, wherein the forest land information extraction in step 3 includes that vegetation forest land is extracted, without vegetation forest land
Extract, forest land is extracted, is extracted to exposed soil and Clean water withdraw, the classification of forest land information and the method extracted are: setting R, G, B difference table
The average value for showing the red channel of image, three green channel, blue channel channels is believed using R, G, B as radix according to different forest lands
Breath sets one-to-one formula to classify to forest land information.
Preferably, vegetation forest land information classifying step is: user-defined feature C, defines such as formula: C=(3 × G
- B-R)/(3 × G+B+R), and if C >=0.238, and 15 ° of gradient <, then to there is the arable land of crop.
Preferably, classification and extraction of the textural characteristics different from the forest land in arable land: formula is set are as follows:Wherein, vk is the sum of the diagonal line of imaged object layer gray level co-occurrence matrixes, if there is vegetation
A >=4.2 in forest land, then to there is forest land in the arable land of crop.
Preferably, forest land information extraction: user-defined feature F, if F is then forest land class in section [35,44] range
Not.
The information extraction of no vegetation forest land and exposed soil ground: setting formula N=((B-G) * (G-R))/255, if N >=4.6,
For no vegetation forest land, after the completion of forest land and forest land information extraction, the more regular brown root tuber of remaining shape is according to its spectrum
Characteristic can extract, the method is as follows: if average value R >=190 of red channel, for exposed soil.
Water body information: in remaining unfiled object, if average brightness >=92, for water body.
Preferably, the data processing of the remote sensing image of the step 2 includes geometric correction, and geometric correction includes following step
It is rapid:
Step (1): selection control point: the place that can obviously position on remote sensing images and topographic map selects control of the same name respectively
Point, to establish the projection relation between image and map, these control points should be selected in the place that can obviously position, as river is handed over
Crunode etc.；
Step (2): it establishes global mapping function: school is determined according to the number of the geometric distortion property of image and ground control point
Positive mathematical model, it is established that the spatial transform relation between image and map, such as Polynomial Method, affine transformation method；
Step (3): resampling interpolation: for corresponding, the root that do not correct image for making the output image picture elements after correction with inputting
According to determining updating formula, to the data permutation of input picture, in resampling, due to the seat of corresponding position calculated
Mark is not integer value, it is necessary to carry out interpolation by the pixel value to surrounding to find out new pixel value.
Preferably, the geometric correction is geometric accurate correction, comprising the following steps:
Step A: the original figure image in input data source；
Step B: it establishes correction transforming function transformation function: using the mapping relations between control point, establishing the transforming function transformation function under least square；
Step C: image output range is determined；
Step D: pixel geometric position transformation；
Step E: the gray scale resampling of pixel；
Step F: being digitized video after output calibration.
Computer can deposit storage medium, be stored with computer program, it is characterised in that: the computer program is by processor
The step of above-described forest resourceies acquisition method based on high definition satellite remote sensing is realized when execution.
Terminal, including processor, it is characterised in that: further include above-described computer readable storage medium, the calculating
The computer program of machine readable storage medium storing program for executing can be executed by processor.
Beneficial effects of the present invention: the forest resourceies acquisition method of the invention based on high definition satellite remote sensing, including it is following
Step: step 1: the first remotely-sensed data source of selection and the second remotely-sensed data source；Step 2: to the first remotely-sensed data source and second
Remotely-sensed data source carries out the data processing of remote sensing image, obtains the first remote sensing image and the second remote sensing image；Step 3: right
First remote sensing image makees forest land information extraction and extracting change information, is used for sampling verification, to form forest land investigation achievement number
According to library；Step 4: extracting change information is made to the first remote sensing image and the second remote sensing image respectively, and will be extracted from the two
Change information out forms Forest road hierarchy information, carries out database update for the database to step 3 to form Forest road hierarchy
Monitor final data, the present invention in, the forest land information extraction in step 3 include vegetation forest land extract, without vegetation
Forest land is extracted, forest land is extracted, extracted to exposed soil and Clean water withdraw, and the classification of forest land information and the method extracted are: setting R, G, B points
Not Biao Shi the red channel of image, three green channel, blue channel channels average value, using R, G, B as radix, according to different woodss
Information setting one-to-one formula in ground classifies for extracting to forest land information, of the invention forest land information classification and mentions
The method taken can be completed rapidly and accurately the classification of forest land information for extracting.
Detailed description of the invention
Invention is described further using attached drawing, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention,
For those of ordinary skill in the art, without creative efforts, it can also be obtained according to the following drawings
Its attached drawing.
Fig. 1 is the flow chart of the forest resourceies acquisition method of the invention based on high definition satellite remote sensing.
Specific embodiment
The invention will be further described with the following Examples.
The present embodiment is related to the forest resourceies acquisition method based on high definition satellite remote sensing, therefore need to obtain suitable remote sensing shadow
Picture, under normal circumstances, there is always some quality problems to carry out making an uproar in necessary processing, such as removal image for remote sensing image
Sound, mist, cloud, shade etc..In order to be conducive to visual interpretation or man computer interactive interpretation, need to do at some necessary Imaging enhanceds
Reason, can protrude forest land information in image, and image can be apparent from, and interpretation property improves.Enhancing processing can be divided into Spectral Characteristic increasing
(prominent grayscale information), space characteristics enhancing (prominent line, edge, texture and structural characteristic) and time information enhancement are (for more by force
For phase).
And be directed to relevant image pre-treatment, in terms of mathematical form, but can be divided into processing (such as linear expansion, ratio,
Histogram transformation etc.) and neighborhood processing (such as convolution algorithm, median filtering, sliding average).
When more phase remote sensing images are interpreted and are classified, Yao Jinhang optimal bands combined.The fusion of remote sensing image is (including not
With the image of resolution ratio, image and merging between other data) interpretation and nicety of grading can be improved.
New wave band (such as vegetation index NDVI or other vegetation indexs) is generated using the original wave band of image, it helps is mentioned
Interpretation property and nicety of grading of high image etc..
The nicety of grading of forestry remote sensing and the remotely-sensed data phase of selection have certain relationship.Remotely-sensed data selection should be according to prison
Area's phenology Meteorological is surveyed, suitable remotely-sensed data source is selected, for Forest road hierarchy monitoring, it may be desirable to which selection is same
The remotely-sensed data of the one same aspect in area, it is possible to reduce, otherwise can be because of the reason of season difference because of the variation that calendar variation generates
And a large amount of false ground class change informations are generated, the extraction of crucial change information can be interfered.If having selected different aspect image numbers
According to, should in image data processing links, dynamic detection algorithm or in assorting process using " penalty method " eliminate image season
Save poor factor.
The forest resourceies acquisition method based on high definition satellite remote sensing of the present embodiment selects the height of the same aspect in areal
Clear remotely-sensed data, as shown in Figure 1, comprising the following steps:
Step 1: the first remotely-sensed data source of selection and the second remotely-sensed data source；
Step 2: the data processing of remote sensing image is carried out to the first remotely-sensed data source and the second remotely-sensed data source, it is distant to obtain first
Feel striograph and the second remote sensing image, the data processing of the remote sensing image includes visual fusion: by the first remotely-sensed data source
Or second the remote sensing image of at least two different times in remotely-sensed data source merged: three wave bands are generated according to multispectral image
Image, Image registration reconnaissance operation before merging, is transformed into HSV image, generates V-band image, and inverse transformation generates RGB image, shadow
Multispectral image spatial resolution not only can be improved as passing through fusion, but also retained its multispectral characteristic.Therefore, it is not only to count
According to simple composite, and emphasize the optimization of information, with the useful thematic information of protrusion, eliminate or inhibit unrelated information, change
The image environment of kind target identification reduces ambiguity (i.e. ambiguity, uncertainty and mistake to enhance the reliability of interpretation
Difference), nicety of grading is improved, application and effect is expanded；
Step 3: making forest land information extraction and extracting change information to the first remote sensing image, sampling verification is used for, to form woods
Investigate raw data in ground；
Step 4: extracting change information is made to the first remote sensing image and the second remote sensing image respectively, and will be extracted from the two
Change information out forms Forest road hierarchy information, carries out database update for the database to step 3 to form Forest road hierarchy
Monitor final data.
Before carrying out the classification of forest land information, Image Segmentation first is carried out to remote sensing image, carries out forest land information point after segmentation again
Class and extraction, specific as follows: the forest land information extraction in step 3 includes that vegetation forest land is extracted, without vegetation forest land
Extract, forest land is extracted, is extracted to exposed soil and Clean water withdraw, the classification of forest land information and the method extracted are: setting R, G, B difference table
The average value for showing the red channel of image, three green channel, blue channel channels is believed using R, G, B as radix according to different forest lands
Breath sets one-to-one formula to classify to forest land information.
Specific classification method is as follows:
Forest land is divided into the arable land for having crop and the arable land without crop, by observing image information and examining or check on the spot, finds without crop
Forest land be brown, have vegetation forest land is presented on image green and yellow, for extraction the category object, customized spy
C is levied, is defined such as formula: C=(3 × G-B-R)/(3 × G+B+R), if C >=0.238, and 15 ° of gradient <, then to there is crop
Arable land.
Using the farmland information for having crop classified, if containing forest land information by the discovery of eye-observation image,
For the forest land information at this time needing these accidentally to be separated to extracting, best mode is exactly to introduce expertise,
That is the general topography in the arable land of long-term cropping is relatively flat, and the forest land for planting forest is usually hills or upward slope, therefore, the gradient
It can be used as the measurement standard for distinguishing the arable land and forest land that have crop, the analysis found that, the gradient is scheduled on less than 15 ° energy
More accurately the plant extraction of crop is come out.
Classification and extraction of the textural characteristics different from the forest land in arable land: formula is set are as follows:Its
In, vk is the sum of the diagonal line of imaged object layer gray level co-occurrence matrixes, if there is A >=4.2 in vegetation forest land, to there is work
Forest land in the arable land of object.
Preferably, the forest land part distribution characteristics in image is more scattered, is sent out by observation image and imaged object
Existing forest land is dark green in image, in order to extract forest land information, user-defined feature F, if F in section [35,44] range,
It is then forest land classification.
No vegetation forest land and naked Land Information: formula N=((B-G) * (G-R))/255 is set, if N >=4.6, for nothing
After the completion of vegetation forest land, forest land and forest land information extraction, the more regular brown root tuber of remaining shape is according to its spectral characteristic
It can extract, the method is as follows: if average value R >=190 of red channel, for exposed soil.
Water body information: in remaining unfiled object, if average brightness >=92, for water body.
After above classification, extract it is certain easily obscure classification when combine human brain cognition introduce expertise, quickly,
It has been accurately finished the information extraction of Type of Forest Land.
The forest land information of the present embodiment is classified and extracting method is compared with the supervised classification of the prior art and unsupervised classification,
It is more simple, it is easy to accomplish, and accuracy rate is higher.
Preferably, the data processing of the remote sensing image of the step 2 includes geometric correction, and remote sensing images are in collection process
In, the variation of remote sensor height and attitude angle, Atmosphere Refraction, earth curvature, hypsography, earth rotation and remote sensor itself are tied
Structure performance etc. can all cause image geometry to deform.Geometry deformation makes the geometric figure in image with the object in selected map
Geometric figure in projection generates difference, and image is made to produce the distortion of geometry or position, is mainly shown as displacement, rotation
Turn, scaling, affine, bending and higher order bending, or shows as picture dot and squeeze, stretching, extension, turn round with respect to the generation of ground physical location
Bent or offset.
It is influenced to eliminate above-mentioned error, so remote sensing image needs to do geometric correction.Geometric correction includes following step
It is rapid:
Step (1): selection control point: the place that can obviously position on remote sensing images and topographic map selects control of the same name respectively
Point, to establish the projection relation between image and map, these control points should be selected in the place that can obviously position, as river is handed over
Crunode etc.；
Step (2): it establishes global mapping function: school is determined according to the number of the geometric distortion property of image and ground control point
Positive mathematical model, it is established that the spatial transform relation between image and map, such as Polynomial Method, affine transformation method；
Step (3): resampling interpolation: for corresponding, the root that do not correct image for making the output image picture elements after correction with inputting
According to determining updating formula, to the data permutation of input picture, in resampling, due to the seat of corresponding position calculated
Mark is not integer value, it is necessary to carry out interpolation by the pixel value to surrounding to find out new pixel value.
The geometric correction method of the present embodiment is compared with the prior art, and the image that can would be more accurately be distorted carries out
It corrects, to go back the original looks of original image.
Preferably, the geometric correction is geometric accurate correction, comprising the following steps:
Step A: the original figure image in input data source；
Step B: it establishes correction transforming function transformation function: using the mapping relations between control point, establishing the transforming function transformation function under least square；
Step C: image output range is determined；
Step D: pixel geometric position transformation；
Step E: the gray scale resampling of pixel；
Step F: being digitized video after output calibration.
When doing geometric accurate correction, control point is evenly distributed as much as possible at 10-30 or so, and last RMS control is in a picture dot
Within.The quantity at control point depend on image resolution ratio and breadth, cannot it is very few can not be excessive.Common straightening die
Type is quadratic polynomial, using can greatly improve calibration result after the geometric accurate correction method of the present embodiment.
The processing of remote sensing image further includes Atmospheric radiation correction.The high registration accuracy between image is carried out, it can be to avoid because of mistake
Position and caused by puppet variation.
Finish geometric correction work after, image is just provided with geodetic coordinates, be provided with specific geographical location, can with it is right
Other maps answered such as topographic map, forest form map, digital elevation (DEM), GPS data etc. do synchronous comparison, assist the differentiation of image.
The present embodiment, which also provides computer, can deposit storage medium, be stored with computer program, and the computer program is processed
The step of device realizes the above-described forest resourceies acquisition method based on high definition satellite remote sensing when executing.
The present embodiment also provides terminal, including processor and above-described computer readable storage medium, the computer
The computer program of readable storage medium storing program for executing can be executed by processor.
Finally it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than the present invention is protected
The limitation of range is protected, although explaining in detail referring to preferred embodiment to the present invention, those skilled in the art are answered
Work as understanding, it can be with modification or equivalent replacement of the technical solution of the present invention are made, without departing from the reality of technical solution of the present invention
Matter and range.
Claims (10)
1. the forest resourceies acquisition method based on high definition satellite remote sensing, it is characterised in that: the following steps are included:
Step 1: the first remotely-sensed data source of selection and the second remotely-sensed data source；
Step 2: the data processing of remote sensing image is carried out to the first remotely-sensed data source and the second remotely-sensed data source, it is distant to obtain first
Feel striograph and the second remote sensing image, the data processing of the remote sensing image includes visual fusion: by the first remotely-sensed data source
Or second the remote sensing image of at least two different times in remotely-sensed data source merged: three wave bands are generated according to multispectral image
Image, Image registration reconnaissance operation before merging, is transformed into HSV image, generates V-band image, and inverse transformation generates RGB image；
Step 3: making forest land information extraction and extracting change information to the first remote sensing image, sampling verification is used for, to form woods
Investigate raw data in ground；
Step 4: extracting change information is made to the first remote sensing image and the second remote sensing image respectively, and will be extracted from the two
Change information out forms Forest road hierarchy information, carries out database update for the database to step 3 to form Forest road hierarchy
Final data is monitored,
Wherein, the forest land information extraction in step 3 includes that vegetation forest land extracts, extracts without vegetation forest land, has woods
Ground is extracted, extracted to exposed soil and Clean water withdraw, and the classification of forest land information and the method extracted are: setting R, G, B, to respectively indicate image red
The average value in channel, three green channel, blue channel channels, using R, G, B as radix, according to different forest land information settings one
One corresponding formula come to forest land information classify for extract.
2. as described in claim 1 based on the forest resourceies acquisition method of high definition satellite remote sensing, it is characterised in that: there is crop to cover
Lid forest land information classifying step is: user-defined feature C, defines such as formula: C=(3 × G-B-R)/(3 × G+B+R), if C >=
0.238, and 15 ° of gradient <, then to there is the arable land of crop.
3. as claimed in claim 2 based on the forest resourceies acquisition method of high definition satellite remote sensing, it is characterised in that: textural characteristics
Different from the classification and extraction of the forest land in arable land: setting formula are as follows:Wherein, vk is imaged object layer ash
The sum of the diagonal line for spending co-occurrence matrix, if there is A >=4.2 in vegetation forest land, to there is forest land in the arable land of crop.
4. as described in claim 1 based on the forest resourceies acquisition method of high definition satellite remote sensing, it is characterised in that: forest land letter
Breath extracts: user-defined feature F, if F is then forest land classification in section [35,44] range.
5. as described in claim 1 based on the forest resourceies acquisition method of high definition satellite remote sensing, it is characterised in that: no crop is covered
The information extraction of lid forest land and exposed soil ground: formula N=((B-G) * (G-R))/255 is set, if N >=4.6, for no vegetation woods
After the completion of ground, forest land and forest land information extraction, the more regular brown root tuber of remaining shape can have been extracted according to its spectral characteristic
Out, the method is as follows: if average value R >=190 of red channel, for exposed soil.
6. as described in claim 1 based on the forest resourceies acquisition method of high definition satellite remote sensing, it is characterised in that: Water-Body Information
It extracts: in remaining unfiled object, if average brightness >=92, for water body.
7. as described in claim 1 based on the forest resourceies acquisition method of high definition satellite remote sensing, it is characterised in that: the step
The data processing of two remote sensing image includes geometric correction, geometric correction the following steps are included:
Step (1): selection control point: the place that can obviously position on remote sensing images and topographic map selects control of the same name respectively
Point, to establish the projection relation between image and map；
Step (2): it establishes global mapping function: school is determined according to the number of the geometric distortion property of image and ground control point
Positive mathematical model, it is established that the spatial transform relation between image and map；
Step (3): resampling interpolation: according to determining updating formula, to the data permutation of input picture, in resampling
In, since the coordinate of corresponding position calculated is not integer value, it is necessary to carry out interpolation by the pixel value to surrounding to find out
New pixel value.
8. as claimed in claim 7 based on the forest resourceies acquisition method of high definition satellite remote sensing, it is characterised in that: the geometry
It is corrected to geometric accurate correction, comprising the following steps:
Step A: the original figure image in input data source；
Step B: it establishes correction transforming function transformation function: using the mapping relations between control point, establishing the transforming function transformation function under least square；
Step C: image output range is determined；
Step D: pixel geometric position transformation；
Step E: the gray scale resampling of pixel；
Step F: being digitized video after output calibration.
9. computer can deposit storage medium, it is stored with computer program, it is characterised in that: the computer program is held by processor
The step of claim 1 to 8 described in any item forest resourceies acquisition methods based on high definition satellite remote sensing are realized when row.
10. terminal, including processor, it is characterised in that: it further include computer readable storage medium as claimed in claim 9, it should
The computer program of computer readable storage medium can be executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910064076.3A CN109977991A (en) | 2019-01-23 | 2019-01-23 | Forest resourceies acquisition method based on high definition satellite remote sensing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910064076.3A CN109977991A (en) | 2019-01-23 | 2019-01-23 | Forest resourceies acquisition method based on high definition satellite remote sensing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109977991A true CN109977991A (en) | 2019-07-05 |
Family
ID=67076751
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910064076.3A Withdrawn CN109977991A (en) | 2019-01-23 | 2019-01-23 | Forest resourceies acquisition method based on high definition satellite remote sensing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109977991A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110502597A (en) * | 2019-08-21 | 2019-11-26 | 湖北泰龙互联通信股份有限公司 | A kind of forest resourceies acquisition method and device based on GIS technology |
| CN111134112A (en) * | 2019-10-28 | 2020-05-12 | 山东省林木种质资源中心 | Forest germplasm resource field collection method and system |
| CN111310614A (en) * | 2020-01-22 | 2020-06-19 | 北京航天宏图信息技术股份有限公司 | Method and device for extracting remote sensing image |
| CN111666827A (en) * | 2020-05-18 | 2020-09-15 | 福建省林业调查规划院(福建省野生动植物与湿地资源监测中心) | Intelligent forestry disease and pest identification method and system |
| CN112561470A (en) * | 2020-12-08 | 2021-03-26 | 海南省林业科学研究院(海南省红树林研究院) | Construction method of digital forestry big data system |
| CN116824396A (en) * | 2023-08-29 | 2023-09-29 | 湖北省泛星信息技术有限公司 | Multi-satellite data fusion automatic interpretation method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103093233A (en) * | 2012-12-03 | 2013-05-08 | 中国环境科学研究院 | Forest classification method based on object-oriented high-resolution remote sensing image |
-
2019
- 2019-01-23 CN CN201910064076.3A patent/CN109977991A/en not_active Withdrawn
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103093233A (en) * | 2012-12-03 | 2013-05-08 | 中国环境科学研究院 | Forest classification method based on object-oriented high-resolution remote sensing image |
Non-Patent Citations (4)
| Title |
|---|
| 王照利 等: "卫星遥感在林地监测中的应用分析", 《林业资源管理》 * |
| 邱春霞 等: "遥感图像几何校正模型探讨", 《安徽农业科学》 * |
| 鲁恒 等: "无人机高空间分辨率影像分类研究", 《测绘科学》 * |
| 黄祚继 等: "《多源遥感数据目标地物的分类与优化》", 31 May 2017, 中国科学技术大学出版社 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110502597A (en) * | 2019-08-21 | 2019-11-26 | 湖北泰龙互联通信股份有限公司 | A kind of forest resourceies acquisition method and device based on GIS technology |
| CN111134112A (en) * | 2019-10-28 | 2020-05-12 | 山东省林木种质资源中心 | Forest germplasm resource field collection method and system |
| CN111310614A (en) * | 2020-01-22 | 2020-06-19 | 北京航天宏图信息技术股份有限公司 | Method and device for extracting remote sensing image |
| CN111310614B (en) * | 2020-01-22 | 2023-07-25 | 航天宏图信息技术股份有限公司 | Remote sensing image extraction method and device |
| CN111666827A (en) * | 2020-05-18 | 2020-09-15 | 福建省林业调查规划院(福建省野生动植物与湿地资源监测中心) | Intelligent forestry disease and pest identification method and system |
| CN111666827B (en) * | 2020-05-18 | 2023-04-18 | 福建省林业调查规划院(福建省野生动植物与湿地资源监测中心) | Forestry disease and pest intelligent identification method and system |
| CN112561470A (en) * | 2020-12-08 | 2021-03-26 | 海南省林业科学研究院(海南省红树林研究院) | Construction method of digital forestry big data system |
| CN116824396A (en) * | 2023-08-29 | 2023-09-29 | 湖北省泛星信息技术有限公司 | Multi-satellite data fusion automatic interpretation method |
| CN116824396B (en) * | 2023-08-29 | 2023-11-21 | 湖北省泛星信息技术有限公司 | Multi-satellite data fusion automatic interpretation method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109977991A (en) | Forest resourceies acquisition method based on high definition satellite remote sensing | |
| CN105740759B (en) | Semilate rice information decision tree classification approach based on feature extraction in multi-temporal data | |
| CN109919875B (en) | High-time-frequency remote sensing image feature-assisted residential area extraction and classification method | |
| CN108764255B (en) | Method for extracting winter wheat planting information | |
| US8712148B2 (en) | Generating agricultural information products using remote sensing | |
| CN108195767B (en) | Estuary wetland foreign species monitoring method | |
| CN108984803B (en) | Method and system for spatializing crop yield | |
| CN107527014A (en) | Crops planting area RS statistics scheme of sample survey design method at county level | |
| Chen et al. | Mapping forest and their spatial–temporal changes from 2007 to 2015 in tropical hainan island by integrating ALOS/ALOS-2 L-Band SAR and landsat optical images | |
| CN111368736A (en) | Rice refined estimation method based on SAR and optical remote sensing data | |
| CN112800973A (en) | Spartina alterniflora extraction method based on vegetation phenological feature decision | |
| CN109671038B (en) | Relative radiation correction method based on pseudo-invariant feature point classification layering | |
| CN109829423A (en) | A kind of icing lake infrared imaging detection method | |
| CN115481368A (en) | Vegetation coverage estimation method based on full remote sensing machine learning | |
| CN105447274A (en) | Method of performing coastal wetland drawing for medium-resolution remote sensing image by utilizing object-oriented classification technology | |
| Wan et al. | Mapping annual urban change using time series Landsat and NLCD | |
| Chen et al. | 3D model construction and ecological environment investigation on a regional scale using UAV remote sensing | |
| Li et al. | Monitoring spatial and temporal patterns of rubber plantation dynamics using time-series landsat images and google earth engine | |
| TAO et al. | Fusing multi-source data to map spatio-temporal dynamics of winter rape on the Jianghan Plain and Dongting Lake Plain, China | |
| Xu et al. | Spatiotemporal distribution of cage and raft aquaculture in China's offshore waters using object-oriented random forest classifier | |
| Zhuang et al. | A Method for water body extraction based on the tasselled cap transformation from remote sensing images | |
| Babykalpana et al. | Classification of land use land cover change detection using remotely sensed data | |
| CN115527108A (en) | Method for rapidly identifying water and soil loss artificial disturbance plots based on multi-temporal Sentinel-2 | |
| Yang et al. | New method for cotton fractional vegetation cover extraction based on UAV RGB images | |
| Bao et al. | A fine digital soil mapping by integrating remote sensing-based process model and deep learning method in Northeast China |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20190705 |
|
| WW01 | Invention patent application withdrawn after publication |