CN110986884A - Unmanned aerial vehicle-based aerial survey data preprocessing and vegetation rapid identification method - Google Patents
Unmanned aerial vehicle-based aerial survey data preprocessing and vegetation rapid identification method Download PDFInfo
- Publication number
- CN110986884A CN110986884A CN201911144743.5A CN201911144743A CN110986884A CN 110986884 A CN110986884 A CN 110986884A CN 201911144743 A CN201911144743 A CN 201911144743A CN 110986884 A CN110986884 A CN 110986884A
- Authority
- CN
- China
- Prior art keywords
- vegetation
- aerial vehicle
- unmanned aerial
- survey data
- aerial survey
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C11/00—Photogrammetry or videogrammetry, e.g. stereogrammetry; Photographic surveying
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C11/00—Photogrammetry or videogrammetry, e.g. stereogrammetry; Photographic surveying
- G01C11/04—Interpretation of pictures
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
- G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
- G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
- G01S19/42—Determining position
- G01S19/43—Determining position using carrier phase measurements, e.g. kinematic positioning; using long or short baseline interferometry
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/188—Vegetation
Abstract
A method for preprocessing aerial survey data and identifying vegetation based on an unmanned aerial vehicle comprises the steps of planning an aerial vehicle flight path to obtain aerial survey data, denoising aerial survey data images of the unmanned aerial vehicle, calculating an overgreen vegetation index of the aerial survey data of the unmanned aerial vehicle, identifying vegetation by using a maximum inter-class variance method based on a genetic algorithm, combining texture features of vegetation in the aerial survey data of the unmanned aerial vehicle, carrying out precision evaluation on vegetation results quickly identified by a system by using visual interpretation pixel by pixel in a man-machine interaction mode, and selecting evaluation indexes to establish an error matrix to calculate overall precision. Utilize unmanned aerial vehicle aerial survey data can be fast, accurate extraction vegetation information, for the balance of maintaining ecosystem provides scientific foundation, and then can realize ecosystem's dynamic balance.
Description
Technical Field
The invention belongs to the technical field of aerial survey remote sensing of unmanned aerial vehicles, and particularly relates to an aerial survey data preprocessing method based on an unmanned aerial vehicle and a vegetation rapid identification system.
Background
Vegetation plays an important role in the "atmospheric-moisture-soil" transformation and also in regulating global climate and global carbon cycle, and is therefore essential for land. At present, the vegetation is monitored by using the traditional method, so that the time and the labor are wasted, the financial resources are wasted, the vegetation is difficult to effectively extract, and particularly, the periodic monitoring is more difficult. The remote sensing technology can acquire data in a large range, can periodically observe, is a common method for monitoring vegetation, and cannot acquire required remote sensing data due to the influence of self time resolution and weather. Manned aerial remote sensing is also limited by weather, environment and flight time.
The unmanned aerial vehicle is based on a wireless remote controller to control the unmanned aerial vehicle to fly on a planned air route. The unmanned aerial vehicle remote sensing platform combines an unmanned aerial vehicle with a photogrammetry system, and quickly acquires space remote sensing information in the aspects of agriculture, forestry, homeland planning and the like. The unmanned aerial vehicle remote sensing can acquire data under the cloud, has higher flexibility ratio, is convenient to take off and land, has short data acquisition period, can realize acquiring multi-scale and high-resolution aerial survey data, can overcome the defects of time and labor waste of the traditional method, and makes up the defects of optical remote sensing and manned aerial remote sensing.
In recent years, along with the pace of urbanization, the expansion of cities causes the ecological environment of cities to be continuously deteriorated, particularly the urban heat island effect. Since 1961, the national average temperature in 2018 in summer has a high history, extreme high temperature exceeding 40 ℃ is frequently generated in Chongqing Fengjie, Wuxi and the like, the temperature of Jilin and Liaoning is extremely high, and the urban ecological environment is increasingly concerned by people. The vegetation is used as the primary productivity in the urban ecological system, can adjust the air temperature and effectively reduce the concentration of PM2.5, has the functions of purifying air, weakening noise, beautifying the environment and the like, and plays an important role in the urban ecological system. Therefore, the method can monitor the vegetation in the city, provide scientific basis for planning, managing, constructing and evaluating the vegetation in the city by various departments, and maintain the stability of the ecological system in the city.
Therefore, the vegetation information of the city can be rapidly and accurately extracted by using the aerial survey data of the unmanned aerial vehicle, a scientific basis is provided for maintaining the balance of the urban ecosystem, and the dynamic balance of the urban ecosystem can be further realized.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a method for preprocessing aerial survey data based on an unmanned aerial vehicle and quickly identifying vegetation, the obtained result is accurate, the method is applied to urban vegetation identification, the difficulties encountered in artificial practice can be reduced, the reliability of vegetation identification is improved, reliable data is provided for departments of urban vegetation monitoring, garden planning and the like, scientific and reliable data basis is provided for maintaining the balance and stability of an urban ecosystem, and further the dynamic balance of the urban ecosystem can be realized
In order to achieve the purpose, the invention adopts the technical scheme that:
the identification method comprises the steps of firstly planning a flight route of the unmanned aerial vehicle to obtain aerial survey data, and obtaining the aerial survey data of the unmanned aerial vehicle. Secondly, carry out the preliminary treatment to unmanned aerial vehicle aerial survey data through unmanned aerial vehicle data processing module, including the image is removed the noise and is handled, and calculates the green vegetation index of crossing of unmanned aerial vehicle aerial survey data, and then carries out binary processing to the image. And thirdly, rapidly identifying the vegetation in the image by using a maximum inter-class variance method based on a genetic algorithm by using a vegetation identification module to obtain an accurate result. And finally, the precision verification module combines the texture characteristics of the vegetation in the original aerial survey data of the unmanned aerial vehicle, adopts a man-machine interaction mode, performs precision evaluation on the vegetation result quickly identified by the system by using visual interpretation pixel by pixel, and selects evaluation indexes to establish an error matrix to calculate the overall precision. The method comprises the following specific steps:
planning a flight route of an unmanned aerial vehicle to obtain aerial survey data to obtain the aerial survey data of the unmanned aerial vehicle;
the system of the unmanned aerial vehicle adopts a ground base station mode, the fuselage carries a GPS RTK device with a 100Hz refresh rate, and the positioning information of the unmanned aerial vehicle with high precision can be obtained in real time while flying, and the positioning information is not only used for automatic navigation and flight attitude adjustment of the unmanned aerial vehicle, but also can ensure that the high precision information is obtained while taking photos by plane. And planning the flight route of the unmanned aerial vehicle through the unmanned aerial vehicle system to acquire aerial survey data, so as to obtain the aerial survey data of the unmanned aerial vehicle.
Secondly, preprocessing the aerial survey data of the unmanned aerial vehicle through an unmanned aerial vehicle image processing module;
1) carrying out image denoising treatment on the aerial survey data of the unmanned aerial vehicle: and (3) adopting mean filtering to weaken Gaussian noise and median filtering to remove salt and pepper noise.
2) And calculating the green vegetation index Exg of the unmanned aerial vehicle aerial survey data for eliminating the noise influence, and performing binarization processing on the image. The over-green vegetation index Exg is calculated using the following formula:
when g > r and g > b, Exg ═ 2 Xg-r-b (1)
Otherwise, Exg is 0
Wherein g is a green band, r is a red band, and b is a blue band.
Thirdly, rapidly identifying vegetation in the image;
and the vegetation identification module is used for segmenting vegetation and non-vegetation in the image after the second step of preprocessing by adopting a maximum inter-class variance method based on a genetic algorithm so as to identify the vegetation in the image.
The solving steps of the maximum inter-class variance method based on the genetic algorithm are as follows:
3.1) since the gray scale value of the image is between 0-255, m random numbers are generated between 0-255 to represent the number of each gray scale value: y is11-Y1m;
3.2) selecting a fitness function based on the maximum inter-class variance method, and classifying vegetation and non-vegetation for the aerial survey data of the unmanned aerial vehicleThe cutting threshold value is recorded as k, and the percentage of the vegetation pixel number in the whole image is recorded as w0Average gray scale of vegetation is recorded as u0And the percentage of the number of the non-vegetation pixels in the whole image is recorded as w1Average intensity of non-vegetation is denoted as u1. The total average gray scale of the aerial survey data of the unmanned aerial vehicle is recorded as u, and the inter-class variance of vegetation and non-vegetation is recorded as G. Let the size of the aerial survey data of the unmanned aerial vehicle be MxN, and the number of pixels with the gray value smaller than the threshold k be recorded as N0The number of pixels with gray value greater than k is recorded as N1Then
w0=N0/(M×N) (2)
w1=N1/(M×N) (3)
N0+N1=M×N (4)
w0+w1=1 (5)
u=u0×w0+u1×w1(6)
G=w0×(u-u0)2+w1×(u-u1)2(7)
Substituting equation (6) into equation (7) may result in a fitness function: g ═ w0×w1×(u0-u1)2(8)
Further, the applicability value of each individual is calculated: f (Y)11)-f(Y1m);
3.3) selecting individuals is to sort the groups according to the fitness and replace the individuals with small fitness by copying the individuals with larger fitness, thereby keeping the excellence of the individuals and further accelerating the convergence. The new population generated was: y is11’-Y1m’。
3.4) random ordering of the order of individuals in the population, for Y11’-Y1mThe individuals in the' are classified pairwise, and are subjected to cross operation according to the cross probability of 0.7 to grow the individuals of two rows and finally form a group Y of the rows11”-Y1m”。
3.5) random pairing of Y according to a variation probability of 0.411”-Y1m"any of the individuals negates the code, thereby producingNew generation of population Y21-Y2m。
3.6) when the difference value of the average fitness of two connected generations is less than 0.01, the evolution tends to be stable; or when the maximum iteration times is reached, the evolution is completed, otherwise, the step 2 is returned.
3.7) taking the individual with the maximum fitness in the population of the last generation as the optimal result sought by the genetic algorithm, namely the corresponding decimal gray value t, and then t is the threshold value obtained by the genetic algorithm.
3.8) the optimal threshold obtained by the genetic algorithm may be the optimal threshold or the quasi-optimal threshold, namely, the quasi-optimal threshold is different from the optimal threshold by about 10%. Because the unmanned aerial vehicle aerial survey data has 256-level gray scale, the difference between the optimal threshold value of vegetation and non-vegetation in the unmanned aerial vehicle aerial survey data and the quasi-optimal threshold value is 25.5. Therefore, in order to ensure that the division threshold of vegetation and non-vegetation of the unmanned aerial vehicle aerial survey data is the optimal solution in the global and local areas, a fluctuation threshold A is set to be 25.5, and local search is carried out by using the maximum inter-class variance method in the range of [ T-A, T + A ], so that the optimal threshold T is obtained.
Fourthly, evaluating the precision based on a precision verification module;
and (3) combining vegetation texture features in original aerial survey data of the unmanned aerial vehicle, performing precision evaluation on vegetation results quickly identified by the system by using visual interpretation pixel by pixel in a man-machine interaction mode, and selecting evaluation indexes to establish an error matrix to calculate overall precision.
The invention has the beneficial effects that: .
(1) The method achieves denoising of the aerial survey data of the unmanned aerial vehicle and vegetation identification of the aerial survey data of the unmanned aerial vehicle by using a maximum inter-class variance method based on a genetic algorithm, adopts a man-machine interaction mode, divides the aerial survey data into vegetation and non-vegetation by pixels through visual interpretation, carries out precision evaluation on vegetation results quickly identified by the system, and improves reliability of estimation results.
(2) The utilization rate of unmanned aerial vehicle aerial survey data is improved, and manpower, material resources, financial resources and the like are saved due to the improved vegetation identification efficiency.
(3) The method can be used for quickly and accurately preprocessing the aerial survey data of the unmanned aerial vehicle, namely identifying the vegetation.
Drawings
Fig. 1 illustrates aerial survey data of an unmanned aerial vehicle according to an embodiment of the present invention;
FIG. 2 illustrates how Gaussian noise and salt and pepper noise are added to aerial survey data of an unmanned aerial vehicle in an embodiment of the present invention;
FIG. 3 is aerial survey data after denoising by an unmanned aerial vehicle in an embodiment of the invention;
fig. 4 is a calculation result of the green vegetation crossing index of the aerial survey data of the unmanned aerial vehicle in the embodiment of the invention;
fig. 5 is a vegetation identification result of aerial survey data of an unmanned aerial vehicle in the embodiment of the present invention;
FIG. 6 is a flow chart of the present invention.
Detailed Description
The present invention is further illustrated by the following specific examples.
An unmanned aerial vehicle-based aerial survey data preprocessing and vegetation identification method comprises the following steps:
planning a flight route of an unmanned aerial vehicle to obtain aerial survey data to obtain the aerial survey data of the unmanned aerial vehicle;
this unmanned aerial vehicle that uses is celestial star unmanned aerial vehicle aerial photography system, and its system adopts ground base station mode, and celestial star unmanned aerial vehicle aerial photography system adopts ground base station mode, is carrying on the GPS RTK equipment of 100Hz refresh rate on the fuselage, can acquire high accuracy unmanned aerial vehicle positioning information in real time when flying in the way, and this positioning information not only is used for unmanned aerial vehicle automatic navigation and flight attitude adjustment, can guarantee to acquire high accuracy information when taking photo by plane moreover. The POS information with fixed solution precision enables the orientation element information with high precision to be arranged among the shooting center, the photo and the ground (ground object) at the moment of shooting of each photo. And acquiring aerial survey data by using the sirius unmanned aerial vehicle according to the planning before flight.
The celestial star unmanned aerial vehicle aerial photography system can make a flight plan according to a survey area range in advance, and meanwhile, the flight plan is created according to survey area requirements, a target area and a desired operation precision value (GSD) which are temporarily specified by operators. The operator can rapidly make flight plans, simulate flights, optimize adjustment and other operations through the clicking range according to actual needs. Under the great condition of topography relief, unmanned aerial vehicle can change flying height according to topography relief self-adaptation, can guarantee unmanned aerial vehicle flight safety and also can guarantee the image overlap degree, ensures the quality that the image acquireed. The celestial star unmanned aerial vehicle aerial photography system adopts a highly integrated one-key design flight plan, and as long as a required operation precision value (GSD) is input, the flight path automatic division, flight height self-adaption, automatic setting of technical parameters such as flight bandwidth and overlapping rate and the like can be realized by importing an operation range file, so that the working efficiency is greatly improved. Photogrammetry and modeling software adopted by the celestial star unmanned aerial vehicle aerial photography system are subjected to aerial three-encryption, high-resolution orthographic images are allowed to be produced, initial values do not need to be set, camera calibration is not needed, and professional-grade aerial photography measurement data are produced through a fully automatic operation process.
In the present embodiment, the practicability of the present invention is verified by taking the national forest park of clean moon pool in vinpocess city, Jilin province and the periphery thereof as an example, as shown in FIG. 1. The Changchun clean moon pool forest park is located in a transition region from Changbai mountain at east of Jilin province to Keerqin grassland at west, the area is more than 90 square kilometers, wherein the area of the water area is 4.3 square kilometers, and the forest coverage rate reaches more than 80 percent, so the park is known as a forest oxygen bar. Meanwhile, the method plays an important role in beautifying urban landscaping and improving ecological environment. Thus, the advantages of the drone can be exploited to provide a system that can quickly identify the plants.
Secondly, preprocessing the aerial survey data of the unmanned aerial vehicle through an unmanned aerial vehicle image processing module;
unmanned aerial vehicle image is when processes such as transmission, storage, owing to receive the influence of gaussian noise and salt and pepper noise, can influence the precision of vegetation identification, therefore need carry out the image to be denoised to the aerial survey data that unmanned aerial vehicle obtained, and then calculate the green vegetation index.
The probability density function of gaussian noise follows gaussian distribution, i.e. normal distribution, the power spectral density is uniformly distributed, but is distributed on each point pixel, and since each point in the image is a pollution point, median filtering cannot be used, and since the mean of the positive-too distribution is zero, the mean filtering can weaken gaussian noise.
The salt and pepper noise is the generated black and white alternating bright and dark point noise, because the salt and pepper noise has approximately equal amplitude values but is randomly distributed on different positions, clean points and polluted points exist in an image, and the median filtering selects proper points to replace the polluted points, so that the salt and pepper noise removing effect is good.
In order to visually verify the image denoising effect, gaussian noise and salt-and-pepper noise with the density of 0.3 are respectively added to an image as shown in fig. 2, and the image denoising effect is further reflected by comparison as shown in fig. 3.
A greenish vegetation index (Exg) is calculated for the drone aerial data that eliminates the effects of noise, as in fig. 4.
When g > r and g > b, Exg ═ 2 Xg-r-b (1)
Otherwise Exg is 0
Wherein g is a green band, r is a red band, and b is a blue band.
Thirdly, rapidly identifying vegetation in the image;
and for the image with the calculated green vegetation index, segmenting vegetation and non-vegetation in the image by using a maximum inter-class variance method based on a genetic algorithm, thereby identifying the vegetation in the image.
3.1) generating m random numbers between 0-255 to represent the number of each gray value: y is11-Y1m;
3.2) calculating the applicability value of each individual: f (Y)11)-f(Y1m):
The fitness function is selected based on a maximum inter-class variance method, for unmanned aerial vehicle aerial survey data, a segmentation threshold value of vegetation and non-vegetation is recorded as k, and the percentage of the number of vegetation pixels in the whole image is recorded as w0Average gray scale of vegetation is recorded as u0And the percentage of the number of the non-vegetation pixels in the whole image is recorded as w1Average intensity of non-vegetation is denoted as u1. The total average gray scale of the aerial survey data of the unmanned aerial vehicle is recorded as u, and the inter-class variance of vegetation and non-vegetation is recorded as G. Setting the size of the aerial survey data of the unmanned aerial vehicle as MxN and the number of pixels with the gray value smaller than a threshold kNumber is N0The number of pixels with gray value greater than k is recorded as N1Then
w0=N0/(M×N) (2)
w1=N1/(M×N) (3)
N0+N1=M×N (4)
w0+w1=1 (5)
u=u0×w0+u1×w1(6)
G=w0×(u-u0)2+w1×(u-u1)2(7)
Substituting (6) into (7) may result in a fitness function as: g ═ w0×w1×(u0-u1)2(8)
Further, the applicability value of each individual is calculated: f (Y)11)-f(Y1m);
3.3) selecting individuals is to sort the groups according to the fitness and replace the individuals with small fitness by copying the individuals with larger fitness, thereby keeping the excellence of the individuals and further accelerating the convergence. The new population generated was: y is11’-Y1m’。
3.4) random ordering of the order of individuals in the population, for Y11’-Y1mThe individuals in the' are classified pairwise, and are subjected to cross operation according to the cross probability of 0.7 to grow the individuals of two rows and finally form a group Y of the rows11”-Y1m”。
3.5) random pairing of Y according to a variation probability of 0.411”-Y1m"any of the individuals in the population is used to invert the code, thereby generating a new generation of population Y21-Y2m。
3.6) when the difference value of the average fitness of two connected generations is less than 0.01, the evolution tends to be stable; or when the maximum iteration times is reached, the evolution is completed, otherwise, the step 2 is returned.
3.7) taking the individual with the maximum fitness in the population of the last generation as the optimal result sought by the genetic algorithm, namely the corresponding decimal gray value t, and then t is the threshold value obtained by the genetic algorithm.
3.8) setting a fluctuation threshold value A to be 25.5, carrying out local search once by using a maximum inter-class variance method within the range of [ T-A, T + A ], so as to obtain an optimal threshold value T, wherein the vegetation identification result of the aerial survey data of the unmanned aerial vehicle is shown in fig. 5.
Fourthly, evaluating the precision based on a precision verification module;
combining the texture characteristics of vegetation in the aerial survey data of the unmanned aerial vehicle, adopting a man-machine interaction mode, utilizing visual interpretation to divide the vegetation into vegetation and background pixel by pixel, carrying out precision evaluation on the vegetation result quickly identified by the system, and selecting evaluation indexes to establish an error matrix to calculate the overall precision.
The specific data are as follows:
the above-mentioned embodiments only express the embodiments of the present invention, but not should be understood as the limitation of the scope of the invention patent, it should be noted that, for those skilled in the art, many variations and modifications can be made without departing from the concept of the present invention, and these all fall into the protection scope of the present invention.
Claims (2)
1. A method for preprocessing aerial survey data and identifying vegetation based on an unmanned aerial vehicle is characterized in that the method comprises the steps of firstly, planning an aerial survey route of the unmanned aerial vehicle to obtain the aerial survey data, and obtaining the aerial survey data of the unmanned aerial vehicle; secondly, preprocessing the aerial survey data of the unmanned aerial vehicle; thirdly, the vegetation identification module rapidly identifies the vegetation in the image by using a maximum inter-class variance method based on a genetic algorithm to obtain an accurate result; finally, carrying out precision evaluation on the vegetation result quickly identified by the system, and selecting evaluation indexes to establish an error matrix to calculate the overall precision; the method comprises the following steps:
firstly, an unmanned aerial vehicle system adopts a ground base station mode, a GPS RTK (global positioning system) equipment with a 100Hz refresh rate is carried on a machine body, and an unmanned aerial vehicle flight path is planned through the unmanned aerial vehicle system to obtain aerial survey data of the unmanned aerial vehicle, so that the aerial survey data of the unmanned aerial vehicle are obtained;
secondly, preprocessing the aerial survey data of the unmanned aerial vehicle obtained in the first step;
1) carrying out image denoising processing on the aerial survey data of the unmanned aerial vehicle;
2) calculating an over-green vegetation index Exg for the unmanned aerial vehicle aerial survey data for eliminating the noise influence, and performing binarization processing on the image; the over-green vegetation index Exg is calculated using the following formula:
when g > r and g > b, Exg ═ 2 Xg-r-b (1)
Otherwise, Exg is 0
Wherein g is a green band, r is a red band, and b is a blue band;
thirdly, segmenting vegetation and non-vegetation in the image preprocessed in the second step by adopting a maximum inter-class variance method based on a genetic algorithm, thereby rapidly identifying the vegetation in the image;
the solving steps of the maximum inter-class variance method based on the genetic algorithm are as follows:
3.1) generating m random numbers between the gray values of the image 0-255, representing the number of each gray value: y is11-Y1m;
3.2) selecting a fitness function based on the maximum inter-class variance method, recording a segmentation threshold value of vegetation and non-vegetation as k and recording the percentage of the number of vegetation pixels in the whole image as w for aerial survey data of the unmanned aerial vehicle0Average gray scale of vegetation is recorded as u0And the percentage of the number of the non-vegetation pixels in the whole image is recorded as w1Average intensity of non-vegetation is denoted as u1(ii) a The total average gray scale of the aerial survey data of the unmanned aerial vehicle is recorded as u, and the inter-class variance of vegetation and non-vegetation is recorded as G; let the size of the unmanned aerial vehicle aerial survey data be MxN, and the number of pixels with the gray value smaller than the threshold k be recorded as N0The number of pixels with gray value greater than k is recorded as N1And then:
w0=N0/(M×N) (2)
w1=N1/(M×N) (3)
N0+N1=M×N (4)
w0+w1=1 (5)
u=u0×w0+u1×w1(6)
G=w0×(u-u0)2+w1×(u-u1)2(7)
substituting equation (6) into equation (7) yields a fitness function: g ═ w0×w1×(u0-u1)2(8)
Further, the applicability value of each individual is calculated: f (Y)11)-f(Y1m);
3.3) selecting individuals is to sort the groups according to the fitness, replace the individuals with small fitness by copying the individuals with larger fitness, keep the excellence of the individuals and further accelerate convergence, and the generated new groups are as follows: y is11’-Y1m’;
3.4) random ordering of the order of individuals in the population, for Y11’-Y1mThe individuals in the' are classified pairwise, and are subjected to cross operation according to the cross probability of 0.7 to grow the individuals of two rows and finally form a group Y of the rows11”-Y1m”;
3.5) random pairing of Y according to a variation probability of 0.411”-Y1m"any of the individuals in the population is used to invert the code, thereby generating a new generation of population Y21-Y2m;
3.6) when the difference value of the average fitness of two connected generations is less than 0.01, the evolution tends to be stable; or when the maximum iteration times is reached, the evolution is finished, otherwise, the step 2 is returned;
3.7) taking the individual with the maximum fitness in the population of the last generation as the optimal result sought by the genetic algorithm, namely the corresponding decimal gray value t, and then t is the threshold value obtained by the genetic algorithm;
3.8) setting a fluctuation threshold value A to be 25.5, and carrying out local search once by using a maximum inter-class variance method in the range of [ T-A, T + A ] to obtain an optimal threshold value T;
and fourthly, combining vegetation texture features in original aerial survey data of the unmanned aerial vehicle, performing precision evaluation on vegetation results quickly identified by the system by using visual interpretation pixel by pixel in a man-machine interaction mode, and selecting evaluation indexes to establish an error matrix to calculate overall precision.
2. The unmanned aerial vehicle aerial survey data preprocessing and vegetation identification method according to claim 1, wherein the image denoising processing on the unmanned aerial vehicle aerial survey data in the second step 1) is: and (3) adopting mean filtering to weaken Gaussian noise and median filtering to remove salt and pepper noise.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911144743.5A CN110986884A (en) | 2019-11-21 | 2019-11-21 | Unmanned aerial vehicle-based aerial survey data preprocessing and vegetation rapid identification method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911144743.5A CN110986884A (en) | 2019-11-21 | 2019-11-21 | Unmanned aerial vehicle-based aerial survey data preprocessing and vegetation rapid identification method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110986884A true CN110986884A (en) | 2020-04-10 |
Family
ID=70085349
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911144743.5A Pending CN110986884A (en) | 2019-11-21 | 2019-11-21 | Unmanned aerial vehicle-based aerial survey data preprocessing and vegetation rapid identification method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110986884A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111784996A (en) * | 2020-07-01 | 2020-10-16 | 思翼科技(深圳)有限公司 | Unmanned aerial vehicle remote control method adaptive to bandwidth |
CN114518104A (en) * | 2022-03-14 | 2022-05-20 | 山东三津房地产评估有限公司 | Territorial surveying and mapping method, system and storage medium based on dynamic remote sensing monitoring technology |
CN116434092A (en) * | 2023-06-14 | 2023-07-14 | 天之翼(苏州)科技有限公司 | Unmanned aerial vehicle aerial survey-based image information analysis method and AI server |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104134221A (en) * | 2014-08-21 | 2014-11-05 | 河海大学 | Image segmentation method based on self-adaptive genetic algorithm and OTSU algorithm |
CN105321166A (en) * | 2015-01-07 | 2016-02-10 | 徐州工程学院 | Annular weld joint edge extraction method based on GAP predictor and self-adaptive genetic algorithm |
CN109583378A (en) * | 2018-11-30 | 2019-04-05 | 东北大学 | A kind of vegetation coverage extracting method and system |
-
2019
- 2019-11-21 CN CN201911144743.5A patent/CN110986884A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104134221A (en) * | 2014-08-21 | 2014-11-05 | 河海大学 | Image segmentation method based on self-adaptive genetic algorithm and OTSU algorithm |
CN105321166A (en) * | 2015-01-07 | 2016-02-10 | 徐州工程学院 | Annular weld joint edge extraction method based on GAP predictor and self-adaptive genetic algorithm |
CN109583378A (en) * | 2018-11-30 | 2019-04-05 | 东北大学 | A kind of vegetation coverage extracting method and system |
Non-Patent Citations (1)
Title |
---|
洪浩 等: "基于最大类间方差法与GA相结合的图像分割算法研究", 《辽宁工业大学学报(自然科学版)》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111784996A (en) * | 2020-07-01 | 2020-10-16 | 思翼科技(深圳)有限公司 | Unmanned aerial vehicle remote control method adaptive to bandwidth |
CN114518104A (en) * | 2022-03-14 | 2022-05-20 | 山东三津房地产评估有限公司 | Territorial surveying and mapping method, system and storage medium based on dynamic remote sensing monitoring technology |
CN116434092A (en) * | 2023-06-14 | 2023-07-14 | 天之翼(苏州)科技有限公司 | Unmanned aerial vehicle aerial survey-based image information analysis method and AI server |
CN116434092B (en) * | 2023-06-14 | 2023-09-26 | 天之翼(苏州)科技有限公司 | Unmanned aerial vehicle aerial survey-based image information analysis method and AI server |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111738165B (en) | Method for extracting individual plant canopy from high-resolution unmanned aerial vehicle visible light remote sensing image | |
CN108109437B (en) | Unmanned aerial vehicle autonomous route extraction and generation method based on map features | |
CN111767865A (en) | Method for inverting mangrove forest biomass by using aerial image and laser data | |
CN110986884A (en) | Unmanned aerial vehicle-based aerial survey data preprocessing and vegetation rapid identification method | |
JP2019520631A (en) | Weed recognition in the natural environment | |
CN107918776A (en) | A kind of plan for land method, system and electronic equipment based on machine vision | |
CN105469098A (en) | Precise LINDAR data ground object classification method based on adaptive characteristic weight synthesis | |
CN114078218B (en) | Adaptive fusion forest smoke and fire identification data augmentation method | |
CN110969654A (en) | Corn high-throughput phenotype measurement method and device based on harvester and harvester | |
Ramli et al. | Homogeneous tree height derivation from tree crown delineation using Seeded Region Growing (SRG) segmentation | |
Gruszczyński et al. | Application of convolutional neural networks for low vegetation filtering from data acquired by UAVs | |
Lyu et al. | Development of phenotyping system using low altitude UAV imagery and deep learning | |
CN113033386B (en) | High-resolution remote sensing image-based transmission line channel hidden danger identification method and system | |
CN112881294B (en) | Unmanned aerial vehicle-based mangrove forest stand health degree evaluation method | |
CN110084120B (en) | Vegetation investigation method, device, system, equipment and medium in island environment | |
CN112084989A (en) | Unmanned aerial vehicle and CNN-based large-range pine wood nematode withered vertical wood intelligent detection method | |
CN116739739A (en) | Loan amount evaluation method and device, electronic equipment and storage medium | |
CN116645321A (en) | Vegetation leaf inclination angle calculation statistical method and device, electronic equipment and storage medium | |
CN110580468A (en) | Single-wood structure parameter extraction method based on image matching point cloud | |
CN115810151A (en) | Deep learning woodland extraction tree species recognition algorithm model | |
CN110070513A (en) | The radiation correction method and system of remote sensing image | |
CN112163639B (en) | Crop lodging grading method based on height distribution feature vector | |
Gruszczyński et al. | Correction of low vegetation impact on UAV-Derived point cloud heights with U-Net networks | |
CN114926732A (en) | Multi-sensor fusion crop deep learning identification method and system | |
CN104251691A (en) | Large-area investigation method for forest gas updating and leaf information and integrated system of method |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200410 |