CN111985445B - Grassland insect pest monitoring system and method based on unmanned aerial vehicle multispectral remote sensing - Google Patents
Grassland insect pest monitoring system and method based on unmanned aerial vehicle multispectral remote sensing Download PDFInfo
- Publication number
- CN111985445B CN111985445B CN202010907338.0A CN202010907338A CN111985445B CN 111985445 B CN111985445 B CN 111985445B CN 202010907338 A CN202010907338 A CN 202010907338A CN 111985445 B CN111985445 B CN 111985445B
- Authority
- CN
- China
- Prior art keywords
- image
- insect
- grassland
- module
- remote sensing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Catching Or Destruction (AREA)
Abstract
The invention belongs to the technical field of grassland pest monitoring, and discloses a grassland pest monitoring system and method based on unmanned aerial vehicle multispectral remote sensing, wherein the grassland pest monitoring system based on unmanned aerial vehicle multispectral remote sensing comprises: the system comprises a remote sensing image acquisition module, a grassland temperature and humidity acquisition module, a main control module, a flight module, a remote sensing image enhancement module, an image characteristic extraction module, a pest analysis module, an early warning module and a display module. According to the invention, the equalization coefficient alpha and the widening coefficient beta are introduced through the remote sensing image enhancement module, and the linear expansion of the gray scale range is simultaneously carried out under the condition of maintaining histogram equalization, so that the purpose of flexibly controlling the image quality of the grassland insects is achieved, and the enhancement effect is good; meanwhile, the insect images of the insect plates are collected through the early warning module, so that the increase of pests in the insect images is determined according to the insect images and a preset comparison image, the threshold range of the number of grassland pests to which the increase of pests belongs is determined, early warning information is generated, and early warning can be performed in time.
Description
Technical Field
The invention belongs to the technical field of grassland pest monitoring, and particularly relates to a grassland pest monitoring system and method based on unmanned aerial vehicle multispectral remote sensing.
Background
At present: in recent years, under the background of factors such as global warming, the population distribution and occurrence rule of insect pests are changed significantly, the insect pest occurrence is in an expanding and aggravating trend, the damage degree to crops, forests and grassland vegetation is increasingly serious, and the safety of human living environment is greatly threatened. The loss caused by the damage of insect pests, the yield reduction of billions of jin of various commercial crops in each year is immeasurable. The monitoring and forecasting of insect pest situation is the key for controlling insect pest, accurate and timely monitoring and forecasting can be performed in advance to control pest spread and transmission, and loss is reduced.
The grassland pests have the advantages of multiple types, wide distribution, rapid propagation and large quantity, and are media for spreading grassland and pasture diseases besides directly causing the loss of grasslands, pastures and products thereof. The harmful insects mainly comprise harmful species of insect class, namely, orthoptera, hemiptera, homoptera, coleoptera, lepidoptera, diptera and the like. However, the grassland insect images acquired by the existing grassland insect pest monitoring system based on unmanned aerial vehicle multispectral remote sensing are not clear; meanwhile, early warning of grassland pests cannot be performed in time.
Through the above analysis, the problems and defects existing in the prior art are as follows: the prior grassland insect image acquired by the grassland insect pest monitoring system based on unmanned aerial vehicle multispectral remote sensing is not clear; meanwhile, early warning of grassland pests cannot be performed in time.
Disclosure of Invention
Aiming at the problems existing in the prior art, the invention provides a grassland insect pest monitoring system and method based on unmanned aerial vehicle multispectral remote sensing.
The invention discloses a grassland insect pest monitoring method based on unmanned aerial vehicle multispectral remote sensing, which comprises the following steps:
step one, collecting a grassland insect image by using a spectrum remote sensing detector on the unmanned aerial vehicle through a remote sensing image collecting module;
step two, collecting grassland environmental temperature and humidity data by using a temperature and humidity sensor through a grassland temperature and humidity collecting module;
controlling each module to work normally by using a master controller through a master control module;
fourthly, performing flight operation by using the unmanned aerial vehicle through a flight module;
step five, the acquired image is enhanced by a remote sensing image enhancement module through an image enhancement program;
the image enhancement processing of the collected image by the remote sensing image enhancement module by utilizing an image enhancement program comprises the following steps: acquiring grassland insect image data of a grassland insect image through a spectrum remote sensing detector on the unmanned aerial vehicle; calculating a histogram of the acquired grassland insect image data; respectively carrying out equalization treatment and stretching treatment on the calculated histogram; synthesizing the histogram after the equalization treatment and the histogram after the widening treatment to obtain an enhanced grassland insect image;
the equalization processing and the widening processing are respectively carried out on the calculated histogram, and the method comprises the following steps:
introducing an equalization coefficient alpha and a broadening coefficient beta, wherein the value range of the equalization coefficient alpha and the broadening coefficient beta is [0,1]; the values of the equilibrium coefficient alpha and the broadening coefficient beta can be determined through manual intervention, and can be automatically determined according to the distribution of the histogram h; the values of the equilibrium coefficient alpha and the broadening coefficient beta are automatically determined, specifically, the alpha and beta values when the evaluation value H (alpha, beta) takes the maximum value; the evaluation value H (alpha, beta) is a function of the quality Q of the treated grassland insect image and the contrast increment value C of the treated grassland insect image;
combining the mathematical pattern g (x, y) =t (f (x, y)) of histogram equalization and the mathematical pattern E (x, y) =e (f (x, y)) of gray scale widening processing, and performing histogram equalization processing and widening processing;
the equalization process is specifically as follows: when the equalization processing is carried out, the range of the value of the equalization coefficient alpha is [0, A ], and according to the formula G1 (x, y) =alpha f (x, y)/A+ (1-alpha/A) G (x, y), a histogram after the equalization adjustment processing is obtained through calculation and synthesis;
when the stretching treatment is carried out, the range of the beta value of the stretching coefficient is [0, B ], and according to the formula G2 (x, y) =beta f (x, y)/B+ (1-beta/B) e (x, y), a histogram after the linear stretching treatment is obtained through calculation and synthesis;
step six, extracting insect characteristic elements in the image by using an image characteristic extraction program through an image characteristic extraction module;
step seven, analyzing the insect attribute by using an insect pest analysis program through an insect pest analysis module;
step eight, the number of pests in the insects is pre-warned by a pre-warning module through a pre-warning program; the display module is used for displaying the collected grassland insect images, the grassland environment temperature and humidity, the analysis result and the early warning information.
In the fifth step, the values of the equalization coefficient α and the widening coefficient β are automatically determined, specifically, the values of α and β when the evaluation value H (α, β) takes the maximum value, which includes:
when the histogram of the processed grassland insect image is uniformly distributed, α is 0; when the histogram of the processed grassland insect image is absolutely unbalanced, α is 1; when the variance or sharpness statistic of the histogram of the processed grassland insect image is minimum and is uniformly distributed, β is 1, and the value of β decreases as the variance or sharpness statistic of the histogram increases.
In the fifth step, the values of the equalization coefficient α and the widening coefficient β are automatically determined according to the following formula:
α＝∑(h(i)–1/L) 2 /(2×(1-1/L))；
β＝1–exp(-∑(h(i)–average(h)) 4 )；
when the difference between h and the uniformly distributed histogram is smaller, namely when the grassland insect image has L gray levels, the occurrence frequency of each gray level is 1/L, alpha takes a small value, otherwise, the value of alpha takes a value close to 1, namely the value of alpha increases along with the increase of the difference between h and the uniformly distributed histogram; when the sharpness of the h distribution is larger, the expansion range is large, namely the value of beta increases along with the increase of the sharpness of h.
In a sixth step, the extracting, by the image feature extracting module, the insect feature element in the image by using the image feature extracting program includes:
1) Constructing an image feature extraction model based on a depth convolution neural network;
2) Training an image feature extraction model;
3) Extracting image feature data of an image to be identified by applying the trained image feature extraction model to obtain the image feature data of the image to be identified;
4) And carrying out similarity analysis on the image characteristic data and the target characteristic data to obtain an identification result.
Further, in step 2), the training of the image feature extraction model includes the following steps:
(1) Carrying out convolution calculation on an image feature extraction model layer by layer based on a depth convolution neural network;
(2) Under the condition that the image feature extraction model completes convolution calculation, classifying image feature data output by a last convolution layer of the image feature extraction model by using a classifier to obtain a classification result;
(3) Determining whether to continue training or end training according to the classification result;
(4) If the training is continued according to the classification result, an optimizer is used for optimizing the image feature extraction model, the optimized image feature extraction model is used as a new image feature extraction model, and convolution calculation is carried out again;
(5) Under the condition that the image feature extraction model completes convolution calculation, classifying image feature data output by a last convolution layer of the image feature extraction model by using a classifier to obtain a classification result;
(6) And determining whether to continue training or end training according to the classification result until the training is ended according to the classification result, so as to obtain a trained image feature extraction model.
In step (1), the image feature data of each input channel of at least one convolution layer in the image feature extraction model based on the depth convolution neural network and the convolution kernels of each input channel are respectively subjected to convolution calculation to obtain the image feature data respectively corresponding to each input channel of the next layer of the convolution layer.
In the eighth step, the early warning module performs early warning on the number of pests in the insects by using an early warning program, including:
step A, carrying out block background reconstruction on an image or video image of a grassland, and adopting self-adaptive updating on an image background;
step B, detecting moving crop pests by using a background difference method, obtaining crop pest images, and collecting insect images of the insect plates according to a preset collection frequency;
step C, determining the increase of pests in the insect image according to the insect image and a preset comparison image; and generating early warning information according to the early warning grade corresponding to the pest increment.
In the step B, the detecting the moving crop pests by using the background difference method, before obtaining the crop pest images according to the preset acquisition frequency, further includes:
presetting a corresponding relation between the increasing amount of the pests and the early warning level.
Further, in the step C, the determining the increasing amount of the pest in the insect image according to the insect image and the preset contrast image includes:
when the insect image is acquired for the first time, taking a blank image of the insect plate as the preset contrast image; generating a binary image of the insect image; image segmentation is carried out on the region with the gray value of 255 in the binary image so as to form a plurality of segmentation regions; the number of the plurality of divided regions is counted to determine the number of grassland pests in the insect image and as an increase amount of the pests.
Another object of the present invention is to provide a grassland pest monitoring system based on unmanned aerial vehicle multispectral remote sensing, which includes:
the remote sensing image acquisition module is connected with the main control module and is used for acquiring a grassland insect image through a spectrum remote sensing detector on the unmanned aerial vehicle;
the grassland temperature and humidity acquisition module is connected with the main control module and used for acquiring grassland environment temperature and humidity data through the temperature and humidity sensor;
the main control module is connected with the remote sensing image acquisition module, the grassland temperature and humidity acquisition module, the flight module, the remote sensing image enhancement module, the image characteristic extraction module, the insect pest analysis module, the early warning module and the display module and used for controlling the normal work of each module through the main controller;
the flight module is connected with the main control module and is used for carrying out flight operation through the unmanned aerial vehicle;
the remote sensing image enhancement module is connected with the main control module and is used for enhancing the acquired image through an image enhancement program;
the image feature extraction module is connected with the main control module and is used for extracting insect feature elements in the image through an image feature extraction program;
the insect pest analysis module is connected with the main control module and used for analyzing insect attributes through an insect pest analysis program;
the early warning module is connected with the main control module and is used for carrying out early warning on the number of pests in the insects through an early warning program;
the display module is connected with the main control module and used for displaying the collected grassland insect images, the grassland environment temperature and humidity, the analysis result and the early warning information through the display.
The invention has the advantages and positive effects that: according to the invention, the equalization coefficient alpha and the widening coefficient beta are introduced through the remote sensing image enhancement module, and the linear expansion of the gray scale range is simultaneously carried out under the condition of maintaining histogram equalization, so that the purpose of flexibly controlling the image quality of the grassland insects is achieved, and the enhancement effect is good; meanwhile, the insect images of the insect plates are collected through the early warning module, so that the increasing amount of pests in the insect images is determined according to the insect images and a preset comparison image, the threshold range of the number of the grassland pests to which the increasing amount of the pests belongs is determined, early warning information is generated according to the grassland pest grade corresponding to the threshold range of the number of the grassland pests, and early warning can be performed in time.
Drawings
Fig. 1 is a flowchart of a grassland pest monitoring method based on unmanned aerial vehicle multispectral remote sensing provided by an embodiment of the invention.
Fig. 2 is a flowchart of extracting insect feature elements in an image by an image feature extraction module using an image feature extraction program according to an embodiment of the present invention.
Fig. 3 is a flowchart for training an image feature extraction model according to an embodiment of the present invention.
Fig. 4 is a flowchart of the early warning of the number of pests in insects by using an early warning program through an early warning module according to an embodiment of the present invention.
Fig. 5 is a block diagram of a grassland pest monitoring system based on multi-spectrum remote sensing of an unmanned aerial vehicle according to an embodiment of the present invention.
In fig. 5: 1. the remote sensing image acquisition module; 2. a grassland temperature and humidity acquisition module; 3. a main control module; 4. a flight module; 5. a remote sensing image enhancement module; 6. an image feature extraction module; 7. a pest analysis module; 8. an early warning module; 9. and a display module.
Detailed Description
The present invention will be described in further detail with reference to the following examples in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Aiming at the problems in the prior art, the invention provides a grassland insect pest monitoring system and method based on unmanned aerial vehicle multispectral remote sensing, and the invention is described in detail below with reference to the accompanying drawings.
As shown in fig. 1, the grassland pest monitoring method based on unmanned aerial vehicle multispectral remote sensing provided by the embodiment of the invention comprises the following steps:
s101, acquiring a grassland insect image by using a spectrum remote sensing detector on the unmanned aerial vehicle through a remote sensing image acquisition module.
S102, collecting grassland environmental temperature and humidity data by using a temperature and humidity sensor through a grassland temperature and humidity collecting module.
S103, the main control module controls each module to work normally by using the main controller.
S104, performing flight operation by using the unmanned aerial vehicle through the flight module.
S105, the acquired image is enhanced by the remote sensing image enhancement module through an image enhancement program.
S106, extracting insect characteristic elements in the image by using an image characteristic extraction program through an image characteristic extraction module.
S107, analyzing the insect attribute by the insect pest analysis module by using the insect pest analysis program.
S108, carrying out early warning on the number of pests in the insects by utilizing an early warning program through an early warning module; the display module is used for displaying the collected grassland insect images, the grassland environment temperature and humidity, the analysis result and the early warning information.
In step S105, the values of the equalization coefficient α and the widening coefficient β provided by the embodiment of the present invention are automatically determined, specifically, the values of α and β when the evaluation value H (α, β) reaches the maximum value include:
when the histogram of the processed grassland insect image is uniformly distributed, α is 0; when the histogram of the processed grassland insect image is absolutely unbalanced, α is 1; when the variance or sharpness statistic of the histogram of the processed grassland insect image is minimum and is uniformly distributed, β is 1, and the value of β decreases as the variance or sharpness statistic of the histogram increases.
In step S105, the values of the equalization coefficient α and the widening coefficient β provided by the embodiment of the present invention are automatically determined according to the following formula:
α＝∑(h(i)–1/L) 2 /(2×(1-1/L))；
β＝1–exp(-∑(h(i)–average(h)) 4 )；
when the difference between h and the uniformly distributed histogram is smaller, namely when the grassland insect image has L gray levels, the occurrence frequency of each gray level is 1/L, alpha takes a small value, otherwise, the value of alpha takes a value close to 1, namely the value of alpha increases along with the increase of the difference between h and the uniformly distributed histogram; when the sharpness of the h distribution is larger, the expansion range is large, namely the value of beta increases along with the increase of the sharpness of h.
As shown in fig. 2, in step S106, the extracting, by the image feature extraction module, of insect feature elements in an image by using an image feature extraction program according to the embodiment of the present invention includes:
s201, constructing an image feature extraction model based on a depth convolution neural network.
S202, training an image feature extraction model.
S203, extracting image feature data of the image to be identified by applying the trained image feature extraction model, and obtaining the image feature data of the image to be identified.
S204, performing similarity analysis on the image characteristic data and the target characteristic data to obtain an identification result.
As shown in fig. 3, in step S202, the training of the image feature extraction model provided by the embodiment of the present invention includes the following steps:
s301, performing convolution calculation on an image feature extraction model based on a depth convolution neural network layer by layer.
S302, under the condition that the image feature extraction model completes convolution calculation, using a classifier to classify the image feature data output by the last convolution layer of the image feature extraction model, and obtaining a classification result.
S303, determining whether to continue training or end training according to the classification result.
And S304, if the continuous training is determined according to the classification result, using an optimizer to optimize the image feature extraction model, taking the optimized image feature extraction model as a new image feature extraction model, and carrying out convolution calculation again.
S305, under the condition that the image feature extraction model completes convolution calculation, using a classifier to classify the image feature data output by the last convolution layer of the image feature extraction model, and obtaining a classification result.
S306, determining whether to continue training or end training according to the classification result until the training is ended according to the classification result, and obtaining a trained image feature extraction model.
In step S301, in the image feature extraction model based on the depth convolutional neural network provided by the embodiment of the present invention, the image feature data of each input channel of at least one convolutional layer and the respective convolution kernels of each input channel are respectively subjected to convolution calculation, so as to obtain the image feature data respectively corresponding to each input channel of the next layer of the convolutional layer.
As shown in fig. 4, in step S108, the early warning module provided by the embodiment of the present invention uses an early warning program to early warn the number of pests in insects, including:
s401, performing block background reconstruction on the image or video image of the grassland, and adopting self-adaptive updating on the image background.
S402, detecting moving crop pests by using a background difference method, acquiring crop pest images, and acquiring insect images of the insect plates according to a preset acquisition frequency.
S403, determining the increase of pests in the insect image according to the insect image and a preset comparison image; and generating early warning information according to the early warning grade corresponding to the pest increment.
In step S402, the method for detecting moving crop pests by using the background difference method according to the embodiment of the present invention further includes, before obtaining the crop pest images according to the preset acquisition frequency:
presetting a corresponding relation between the increasing amount of the pests and the early warning level.
In step S403, determining an increase amount of pests in the insect image according to the insect image and the preset contrast image provided in the embodiment of the present invention includes:
when the insect image is acquired for the first time, taking a blank image of the insect plate as the preset contrast image; generating a binary image of the insect image; image segmentation is carried out on the region with the gray value of 255 in the binary image so as to form a plurality of segmentation regions; the number of the plurality of divided regions is counted to determine the number of grassland pests in the insect image and as an increase amount of the pests.
As shown in fig. 5, a grassland pest monitoring system based on unmanned aerial vehicle multispectral remote sensing provided by an embodiment of the invention includes:
the remote sensing image acquisition module 1 is connected with the main control module 3 and is used for acquiring a grassland insect image through a spectrum remote sensing detector on the unmanned aerial vehicle;
the grassland temperature and humidity acquisition module 2 is connected with the main control module 3 and is used for acquiring grassland environment temperature and humidity data through a temperature and humidity sensor;
the main control module 3 is connected with the remote sensing image acquisition module 1, the grassland temperature and humidity acquisition module 2, the flight module 4, the remote sensing image enhancement module 5, the image feature extraction module 6, the insect pest analysis module 7, the early warning module 8 and the display module 9 and is used for controlling the normal work of each module through the main controller;
the flight module 4 is connected with the main control module 3 and is used for carrying out flight operation through the unmanned aerial vehicle;
the remote sensing image enhancement module 5 is connected with the main control module 3 and is used for enhancing the acquired image through an image enhancement program;
the image feature extraction module 6 is connected with the main control module 3 and is used for extracting insect feature elements in the image through an image feature extraction program;
the insect pest analysis module 7 is connected with the main control module 3 and is used for analyzing insect attributes through an insect pest analysis program;
the early warning module 8 is connected with the main control module 3 and is used for carrying out early warning on the number of pests in the insects through an early warning program;
the display module 9 is connected with the main control module 3 and is used for displaying the collected grassland insect images, the grassland environment temperature and humidity, the analysis result and the early warning information through a display.
When the method works, firstly, a remote sensing image acquisition module 1 is used for acquiring a grassland insect image by utilizing a spectrum remote sensing detector on an unmanned plane; the grassland temperature and humidity acquisition module 2 is used for acquiring grassland environmental temperature and humidity data by using a temperature and humidity sensor; secondly, the main control module 3 performs flight operation by using an unmanned aerial vehicle through the flight module 4; the acquired image is enhanced by the remote sensing image enhancement module 5 by utilizing an image enhancement program; extracting insect characteristic elements in the image by using an extraction program through an image characteristic extraction module 6; analyzing the insect attribute by the insect pest analysis module 7 using an analysis program; then, the number of pests in the insects is pre-warned by a pre-warning module 8 through a pre-warning program; finally, the collected grassland insect images, the grassland environment temperature and humidity, the analysis result and the early warning information are displayed by a display module 9 through a display.
In the description of the present invention, unless otherwise indicated, the meaning of "a plurality" is two or more; the terms "upper," "lower," "left," "right," "inner," "outer," "front," "rear," "head," "tail," and the like are used as an orientation or positional relationship based on that shown in the drawings, merely to facilitate description of the invention and to simplify the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore should not be construed as limiting the invention. Furthermore, the terms "first," "second," "third," and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
The foregoing is merely illustrative of specific embodiments of the present invention, and the scope of the invention is not limited thereto, but any modifications, equivalents, improvements and alternatives falling within the spirit and principles of the present invention will be apparent to those skilled in the art within the scope of the present invention.
Claims (10)
1. The grassland insect pest monitoring method based on the unmanned aerial vehicle multispectral remote sensing is characterized by comprising the following steps of:
step one, collecting a grassland insect image by using a spectrum remote sensing detector on the unmanned aerial vehicle through a remote sensing image collecting module;
step two, collecting grassland environmental temperature and humidity data by using a temperature and humidity sensor through a grassland temperature and humidity collecting module;
controlling each module to work normally by using a master controller through a master control module;
fourthly, performing flight operation by using the unmanned aerial vehicle through a flight module;
step five, the acquired image is enhanced by a remote sensing image enhancement module through an image enhancement program;
the image enhancement processing of the collected image by the remote sensing image enhancement module by utilizing an image enhancement program comprises the following steps: acquiring grassland insect image data of a grassland insect image through a spectrum remote sensing detector on the unmanned aerial vehicle; calculating a histogram of the acquired grassland insect image data; respectively carrying out equalization treatment and stretching treatment on the calculated histogram; synthesizing the histogram after the equalization treatment and the histogram after the widening treatment to obtain an enhanced grassland insect image;
the equalization processing and the widening processing are respectively carried out on the calculated histogram, and the method comprises the following steps:
introducing an equalization coefficient alpha and a broadening coefficient beta, wherein the value range of the equalization coefficient alpha and the broadening coefficient beta is [0,1]; the values of the equilibrium coefficient alpha and the broadening coefficient beta can be determined through manual intervention, and can be automatically determined according to the distribution of the histogram h; the values of the equilibrium coefficient alpha and the broadening coefficient beta are automatically determined, specifically, the alpha and beta values when the evaluation value H (alpha, beta) takes the maximum value; the evaluation value H (alpha, beta) is a function of the quality Q of the treated grassland insect image and the contrast increment value C of the treated grassland insect image;
combining the mathematical pattern g (x, y) =t (f (x, y)) of histogram equalization and the mathematical pattern E (x, y) =e (f (x, y)) of gray scale widening processing, and performing histogram equalization processing and widening processing;
the equalization process is specifically as follows: when the equalization processing is carried out, the range of the value of the equalization coefficient alpha is [0, A ], and according to the formula G1 (x, y) =alpha f (x, y)/A+ (1-alpha/A) G (x, y), a histogram after the equalization adjustment processing is obtained through calculation and synthesis;
when the stretching treatment is carried out, the range of the beta value of the stretching coefficient is [0, B ], and according to the formula G2 (x, y) =beta f (x, y)/B+ (1-beta/B) e (x, y), a histogram after the linear stretching treatment is obtained through calculation and synthesis;
step six, extracting insect characteristic elements in the image by using an image characteristic extraction program through an image characteristic extraction module;
step seven, analyzing the insect attribute by using an insect pest analysis program through an insect pest analysis module;
step eight, the number of pests in the insects is pre-warned by a pre-warning module through a pre-warning program; the display module is used for displaying the collected grassland insect images, the grassland environment temperature and humidity, the analysis result and the early warning information.
2. The grassland insect pest monitoring method based on unmanned aerial vehicle multispectral remote sensing as set forth in claim 1, wherein in the fifth step, the values of the equalization coefficient α and the widening coefficient β are automatically determined, specifically, the α and β values when the evaluation value H (α, β) takes the maximum value, include:
when the histogram of the processed grassland insect image is uniformly distributed, α is 0; when the histogram of the processed grassland insect image is absolutely unbalanced, α is 1; when the variance or sharpness statistic of the histogram of the processed grassland insect image is minimum and is uniformly distributed, β is 1, and the value of β decreases as the variance or sharpness statistic of the histogram increases.
3. The grassland insect pest monitoring method based on unmanned aerial vehicle multispectral remote sensing as set forth in claim 1, wherein in the fifth step, the values of the equalization coefficient alpha and the widening coefficient beta are automatically determined according to the following formula:
α＝∑(h(i)–1/L) 2 /(2×(1-1/L))；
β＝1–exp(-∑(h(i)–average(h)) 4 )；
when the difference between h and the uniformly distributed histogram is smaller, namely when the grassland insect image has L gray levels, the occurrence frequency of each gray level is 1/L, alpha takes a small value, otherwise, the value of alpha takes a value close to 1, namely the value of alpha increases along with the increase of the difference between h and the uniformly distributed histogram; when the sharpness of the h distribution is larger, the expansion range is large, namely the value of beta increases along with the increase of the sharpness of h.
4. The grassland insect pest monitoring method based on unmanned aerial vehicle multispectral remote sensing as set forth in claim 1, wherein in step six, the extracting insect characteristic elements in the image by the image characteristic extracting module using the image characteristic extracting program comprises:
1) Constructing an image feature extraction model based on a depth convolution neural network;
2) Training an image feature extraction model;
3) Extracting image feature data of an image to be identified by applying the trained image feature extraction model to obtain the image feature data of the image to be identified;
4) And carrying out similarity analysis on the image characteristic data and the target characteristic data to obtain an identification result.
5. The grassland insect pest monitoring method based on unmanned aerial vehicle multispectral remote sensing as set forth in claim 4, wherein in step 2), the training of the image feature extraction model comprises the following steps:
(1) Carrying out convolution calculation on an image feature extraction model layer by layer based on a depth convolution neural network;
(2) Under the condition that the image feature extraction model completes convolution calculation, classifying image feature data output by a last convolution layer of the image feature extraction model by using a classifier to obtain a classification result;
(3) Determining whether to continue training or end training according to the classification result;
(4) If the training is continued according to the classification result, an optimizer is used for optimizing the image feature extraction model, the optimized image feature extraction model is used as a new image feature extraction model, and convolution calculation is carried out again;
(5) Under the condition that the image feature extraction model completes convolution calculation, classifying image feature data output by a last convolution layer of the image feature extraction model by using a classifier to obtain a classification result;
(6) And determining whether to continue training or end training according to the classification result until the training is ended according to the classification result, so as to obtain a trained image feature extraction model.
6. The grassland insect pest monitoring method based on unmanned aerial vehicle multispectral remote sensing according to claim 5, wherein in the step (1), the image feature data of each input channel of at least one convolution layer in the image feature extraction model based on the depth convolution neural network and the convolution kernels of each input channel are respectively subjected to convolution calculation to obtain the image feature data respectively corresponding to each input channel of the next layer of the convolution layer.
7. The grassland pest monitoring method based on unmanned aerial vehicle multispectral remote sensing as set forth in claim 1, wherein in the eighth step, the early warning module is used for early warning the number of pests in insects by using an early warning program, and the method comprises the following steps:
step A, carrying out block background reconstruction on an image or video image of a grassland, and adopting self-adaptive updating on an image background;
step B, detecting moving crop pests by using a background difference method, obtaining crop pest images, and collecting insect images of the insect plates according to a preset collection frequency;
step C, determining the increase of pests in the insect image according to the insect image and a preset comparison image; and generating early warning information according to the early warning grade corresponding to the pest increment.
8. The method for monitoring plant pests based on multi-spectral remote sensing of unmanned aerial vehicle according to claim 7, wherein in the step B, before the moving crop pests are detected by the background differentiation method and the crop pest images are obtained according to the preset acquisition frequency, the method further comprises:
presetting a corresponding relation between the increasing amount of the pests and the early warning level.
9. The method for monitoring plant insect pests based on multispectral remote sensing of unmanned aerial vehicle according to claim 7, wherein in the step C, the determining the increase of the insect pests in the insect image according to the insect image and the preset contrast image comprises:
when the insect image is acquired for the first time, taking a blank image of the insect plate as the preset contrast image; generating a binary image of the insect image; image segmentation is carried out on the region with the gray value of 255 in the binary image so as to form a plurality of segmentation regions; the number of the plurality of divided regions is counted to determine the number of grassland pests in the insect image and as an increase amount of the pests.
10. A grassland pest monitoring system based on unmanned aerial vehicle multispectral remote sensing applying the grassland pest monitoring method based on unmanned aerial vehicle multispectral remote sensing of any one of claims 1 to 9, characterized in that the grassland pest monitoring system based on unmanned aerial vehicle multispectral remote sensing comprises:
the remote sensing image acquisition module is connected with the main control module and is used for acquiring a grassland insect image through a spectrum remote sensing detector on the unmanned aerial vehicle;
the grassland temperature and humidity acquisition module is connected with the main control module and used for acquiring grassland environment temperature and humidity data through the temperature and humidity sensor;
the main control module is connected with the remote sensing image acquisition module, the grassland temperature and humidity acquisition module, the flight module, the remote sensing image enhancement module, the image characteristic extraction module, the insect pest analysis module, the early warning module and the display module and used for controlling the normal work of each module through the main controller;
the flight module is connected with the main control module and is used for carrying out flight operation through the unmanned aerial vehicle;
the remote sensing image enhancement module is connected with the main control module and is used for enhancing the acquired image through an image enhancement program;
the image feature extraction module is connected with the main control module and is used for extracting insect feature elements in the image through an image feature extraction program;
the insect pest analysis module is connected with the main control module and used for analyzing insect attributes through an insect pest analysis program;
the early warning module is connected with the main control module and is used for carrying out early warning on the number of pests in the insects through an early warning program;
the display module is connected with the main control module and used for displaying the collected grassland insect images, the grassland environment temperature and humidity, the analysis result and the early warning information through the display.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010907338.0A CN111985445B (en) | 2020-09-02 | 2020-09-02 | Grassland insect pest monitoring system and method based on unmanned aerial vehicle multispectral remote sensing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010907338.0A CN111985445B (en) | 2020-09-02 | 2020-09-02 | Grassland insect pest monitoring system and method based on unmanned aerial vehicle multispectral remote sensing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111985445A CN111985445A (en) | 2020-11-24 |
| CN111985445B true CN111985445B (en) | 2023-05-30 |
Family
ID=73446814
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010907338.0A Active CN111985445B (en) | 2020-09-02 | 2020-09-02 | Grassland insect pest monitoring system and method based on unmanned aerial vehicle multispectral remote sensing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111985445B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113189015A (en) * | 2021-04-23 | 2021-07-30 | 宿州学院 | Hyperspectral remote sensing-based insect pest monitoring system and method |
| CN113378700B (en) * | 2021-06-08 | 2022-11-08 | 四川农业大学 | Grassland rat damage dynamic monitoring method based on unmanned aerial vehicle aerial image |
| CN116029430A (en) * | 2022-12-27 | 2023-04-28 | 江苏师范大学科文学院 | Grassland ecological environment monitoring system based on aerial image |
| CN115879773B (en) * | 2023-02-22 | 2023-05-16 | 广东省农业科学院植物保护研究所 | Intelligent disease and pest early warning method and system based on Internet of things |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102750684A (en) * | 2012-06-07 | 2012-10-24 | 邝君 | Method, program and device for image enhancement |
| CN104881865A (en) * | 2015-04-29 | 2015-09-02 | 北京林业大学 | Forest disease and pest monitoring and early warning method and system based on unmanned plane image analysis |
| CN106650750A (en) * | 2016-11-21 | 2017-05-10 | 云南电网有限责任公司电力科学研究院 | UAV remote sensing image processing system for transmission line geological hazard detection |
| CN111460990A (en) * | 2020-03-31 | 2020-07-28 | 青海省草原总站 | Big data-based alpine pastoral area grassland insect pest monitoring and early warning system and method |
-
2020
- 2020-09-02 CN CN202010907338.0A patent/CN111985445B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102750684A (en) * | 2012-06-07 | 2012-10-24 | 邝君 | Method, program and device for image enhancement |
| CN104881865A (en) * | 2015-04-29 | 2015-09-02 | 北京林业大学 | Forest disease and pest monitoring and early warning method and system based on unmanned plane image analysis |
| CN106650750A (en) * | 2016-11-21 | 2017-05-10 | 云南电网有限责任公司电力科学研究院 | UAV remote sensing image processing system for transmission line geological hazard detection |
| CN111460990A (en) * | 2020-03-31 | 2020-07-28 | 青海省草原总站 | Big data-based alpine pastoral area grassland insect pest monitoring and early warning system and method |
Non-Patent Citations (1)
| Title |
|---|
| 无人机技术在草原生态遥感监测中的应用与探讨;李风贤;;测绘通报(第07期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111985445A (en) | 2020-11-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111985445B (en) | Grassland insect pest monitoring system and method based on unmanned aerial vehicle multispectral remote sensing | |
| US10747999B2 (en) | Methods and systems for pattern characteristic detection | |
| CN109308697B (en) | Leaf disease identification method based on machine learning algorithm | |
| Oo et al. | Plant leaf disease detection and classification using image processing | |
| Francis et al. | Identification of leaf diseases in pepper plants using soft computing techniques | |
| Es-saady et al. | Automatic recognition of plant leaves diseases based on serial combination of two SVM classifiers | |
| CN110717903A (en) | Method for detecting crop diseases by using computer vision technology | |
| CN109325431B (en) | Method and device for detecting vegetation coverage in feeding path of grassland grazing sheep | |
| CN112101159B (en) | Multi-temporal forest remote sensing image change monitoring method | |
| KR20160063128A (en) | Apparatus and Method for Computer Aided Diagnosis | |
| WO2018120634A1 (en) | Method and apparatus for identifying disease and insect damage | |
| CN116229335A (en) | Livestock and poultry farm environment recognition method based on image data | |
| CN116958572B (en) | Leaf disease and pest area analysis method in fruit tree breeding | |
| CN115861721B (en) | Livestock and poultry breeding spraying equipment state identification method based on image data | |
| CN114818909A (en) | Weed detection method and device based on crop growth characteristics | |
| Zhang et al. | Evaluation of efficacy of fungicides for control of wheat fusarium head blight based on digital imaging | |
| CN108460370B (en) | Fixed poultry life information alarm device | |
| Narisetti et al. | Automated spike detection in diverse european wheat plants using textural features and the frangi filter in 2d greenhouse images | |
| Dwivedi et al. | Study of machine learning techniques for plant disease recognition in agriculture | |
| Kumar et al. | Automatic leaf disease detection and classification using hybrid features and supervised classifier | |
| CN115965875B (en) | Intelligent monitoring method and system for crop diseases and insect pests | |
| CN114170673A (en) | Method for identifying pig feeding behavior in video based on convolutional neural network | |
| Mladenov et al. | Quality assessment of grain samples using color image analysis | |
| CN114022831A (en) | Binocular vision-based livestock body condition monitoring method and system | |
| CN111160252A (en) | Artificial intelligence-based potato disease monitoring method and system and electronic equipment |
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 |