CN108764284B - Classification and denoising method and system for high-resolution image of dead pine - Google Patents
Classification and denoising method and system for high-resolution image of dead pine Download PDFInfo
- Publication number
- CN108764284B CN108764284B CN201810364964.2A CN201810364964A CN108764284B CN 108764284 B CN108764284 B CN 108764284B CN 201810364964 A CN201810364964 A CN 201810364964A CN 108764284 B CN108764284 B CN 108764284B
- Authority
- CN
- China
- Prior art keywords
- disease
- denoising
- target pixel
- tree
- classification
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
Abstract
The invention discloses a classification denoising method and a classification denoising system for a high-resolution image of a dead pine tree with a disease and pest damage, wherein corresponding feature color feature extraction is carried out on all pixel points in an image sample to be classified through classification training of feature color features according to verified pest and disease tree spectrum information; then screening out a target pixel set which accords with the characteristics of the multi-stage disease pine from the image sample to be classified after the ground object color characteristics are extracted; carrying out primary denoising on the target pixel set conforming to the characteristics of the multilevel disease tree; secondly, performing secondary denoising on the target pixel set conforming to the characteristics of the multilevel disease trees according to the distribution characteristics of the disease trees; and finally, carrying out three-time denoising on the target pixel set conforming to the characteristics of the multistage diseased trees according to the forest region background, and finally generating a classification result of the pine tree disease dead tree state, thereby realizing high detection precision identification of different disease areas and different disease degrees on the high-resolution image of the pine tree disease dead tree, and having high processing speed and high stability.
Description
Technical Field
The invention relates to a pine diseased dead wood identification method, in particular to a classification denoising method and a classification denoising system for a high-resolution image of a pine diseased dead wood.
Background
The spectral feature method is the most common pine identification method for diseases and pests, and is based on the fact that when plants are impregnated by diseases and pests, the difference of physiological changes is reflected on spectral characteristics, particularly the difference of the spectral characteristics of a red region and a near infrared region. The spectral data information of reflection and radiation is obtained by using imaging equipment such as a spectrometer, and the change rule of the reflection spectrum of different wave bands is found on the basis.
Compared with the traditional manual field investigation method, the spectral characteristic information of all the objects in the remote sensing image is obtained to identify the damaged pine, and the method has the advantages of high identification rate, wide identification range, high identification speed and the like. However, due to the influence of sensor resolution, satellite operation cycle and atmospheric environment (such as illumination, temperature, etc.), the spectral feature data of the ground features acquired by an imaging device such as a high-speed spectrometer cannot well characterize the types of the ground features.
From the perspective of image processing and pattern recognition, the description using the support vector data is also a method for identifying diseased pine trees. The method comprises the following basic steps: firstly, extracting color components as color features of corresponding pixel points according to the characteristics of different ground features, and then constructing a weighted support vector data description multi-classification model by establishing a weight function related to the center distance of a training sample so as to realize classification identification of damaged pines.
The support vector data description multi-classification method carries out classification and identification based on image pixels, and has the disadvantages that a large number of samples need to be trained, so that the execution efficiency is low; furthermore, this method does not allow identification of multi-stage diseased pine trees, such as early stage infection, mid-stage infection, late stage infection and dead pine trees; in addition, the method has the defects of high misjudgment rate, relatively low identification precision and the like when solving the multi-classification identification problem.
Disclosure of Invention
In view of the above technical problems, the present invention aims to: the method and the system for efficiently and accurately classifying and denoising the high-resolution images of the dead pine trees with different disease areas and different disease degrees are provided.
The invention provides a classification denoising method for a high-resolution image of a dead pine tree with a pine tree disease, which comprises the following steps:
s1, performing corresponding ground object color feature extraction on all pixel points in an image sample to be classified according to classification training of ground object color features of verified pest tree spectrum information;
s2, screening out a target pixel set which accords with the characteristics of the multi-level diseased pine from the image sample to be classified after the ground object color characteristics are extracted;
s3, carrying out primary denoising on the target pixel set conforming to the characteristics of the multilevel disease tree, and removing misjudged single points or sporadic points from the target pixel set conforming to the characteristics of the multilevel disease tree;
s4, carrying out secondary denoising on the target pixel set conforming to the multi-stage disease tree characteristics according to the disease tree distribution characteristics, and removing large-area noise points in non-forest regions;
and S5, carrying out three-time denoising on the target pixel set conforming to the characteristics of the multilevel disease tree according to the forest region background, and avoiding the misjudgment condition similar to the forest region edge.
The invention provides a classification denoising system for high-resolution images of dead pine tree diseases, which comprises:
the color feature classification and extraction module is used for performing corresponding feature color feature extraction on all pixel points in the image sample to be classified according to the classification training of the feature color features of the verified pest tree spectrum information;
the pixel screening module is used for screening a target pixel set which accords with the characteristics of the multi-level disease pine from the image sample to be classified after the ground object color characteristics are extracted;
the primary denoising module is used for carrying out primary denoising on the target pixel set conforming to the multi-level disease tree characteristics and removing misjudged single points or sporadic points from the target pixel set conforming to the multi-level disease tree characteristics;
the secondary denoising module is used for carrying out secondary denoising on the target pixel set conforming to the multi-stage disease tree characteristics according to the disease tree distribution characteristics and removing large-area noise points of non-forest regions;
and the third denoising module is used for carrying out three times of denoising on the target pixel set conforming to the characteristics of the multilevel disease trees according to the forest region background so as to avoid the misjudgment condition similar to the forest region edge.
The invention relates to a classification denoising method and a classification denoising system for high-resolution images of dead pine diseases, which are used for extracting corresponding ground object color features of all pixel points in an image sample to be classified through classification training of the ground object color features according to verified spectrum information of disease and insect pests trees; then screening out a target pixel set which accords with the characteristics of the multi-stage disease pine from the image sample to be classified after the ground object color characteristics are extracted; carrying out primary denoising on the target pixel set conforming to the characteristics of the multilevel disease tree, and removing misjudged single points or sporadic points from the target pixel set conforming to the characteristics of the multilevel disease tree; secondly, carrying out secondary denoising on the target pixel set conforming to the characteristics of the multilevel disease trees according to the distribution characteristics of the disease trees, and removing large-area noise points in non-forest regions; and finally, carrying out three-time denoising on the target pixel set conforming to the characteristics of the multistage diseased trees according to the forest region background, avoiding the misjudgment condition similar to the edge of the forest region, and finally generating a classification result of the pine tree disease dead tree state, thereby realizing high detection precision identification of different disease areas and different disease degrees on the high-resolution image of the pine tree disease dead tree, and having high processing speed and extremely high stability.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required to be used in the embodiments or the prior art description will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings may be obtained according to these drawings without inventive labor.
Fig. 1 is a flowchart of a classification denoising method for a high-resolution image of a dead pine tree with a pine tree disease according to an embodiment of the present invention;
fig. 2 is a sample of an image to be classified obtained after preprocessing raw image data according to an embodiment of the present invention;
fig. 3 is a block diagram of a sub-flow of step S2 according to an embodiment of the present invention;
FIG. 4 is a block diagram of a sub-flow of step S3 provided by an embodiment of the present invention;
FIG. 5 is a schematic diagram illustrating a location of a cluster center identified in an image file according to an embodiment of the present invention;
FIG. 6 is a table comparing the classification results with the ground surface real information according to the embodiment of the present invention;
FIG. 7 is a block diagram of a classification denoising system for high-resolution images of dead pine tree diseases according to an embodiment of the present invention;
FIG. 8 is a block diagram of the pixel screening module according to an embodiment of the present invention;
fig. 9 is a block diagram of a unit of a primary denoising module according to an embodiment of the present invention.
Detailed Description
In order to make the objects, features and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the embodiments described below are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular system structures, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.
It will be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
It is also to be understood that the terminology used in the description of the present application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in the specification of the present application and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
It should be further understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items. As used in this specification and the appended claims, the term "if" may be interpreted contextually as "when", "upon" or "in response to a determination" or "in response to a detection". Similarly, the phrase "if it is determined" or "if a [ described condition or event ] is detected" may be interpreted contextually to mean "upon determining" or "in response to determining" or "upon detecting [ described condition or event ]" or "in response to detecting [ described condition or event ]".
As shown in fig. 1, the method for classifying and denoising a high-resolution image of a dead pine tree with a pine tree disease provided in an embodiment of the present invention includes:
s1, performing corresponding ground object color feature extraction on all pixel points in an image sample to be classified according to classification training of ground object color features of verified pest tree spectrum information.
The method for classifying and denoising the high-resolution image of the dead pine is realized based on a Microsoft Visual Studio 2008 software platform. Specifically, first, the original image data is preprocessed by geometric correction, radiation correction, atmospheric correction, and other methods, and the processed image result is shown in fig. 2;
firstly, converting DN value of original image data into radiance or atmospheric outer layer surface reflectivity, and eliminating errors generated by a sensor; then, the radiation brightness or the apparent reflectivity is converted into the actual reflectivity of the earth surface so as to eliminate errors caused by atmospheric scattering, absorption and reflection; and finally, the image data of the same area obtained at different time and different wave bands are subjected to geometric transformation to ensure that the image points with the same name are completely superposed on the position and the direction to form an image sample to be classified.
And then, respectively extracting the feature color features of the ground features from the verified spectrum information of the pest trees, classifying the feature color features of the ground features according to different stage states of the dead pine tree diseases, and generating a training sample and a plurality of parameter rules.
Specifically, the method is divided into the following steps according to the state of the pine tree: the method comprises the following steps of extracting the color features of ground objects of verified pest tree spectral information respectively by four categories of early infection, middle infection, later infection and dead pine trees, classifying the color features of the ground objects according to different stage states of dead pine tree diseases, generating a training sample and a plurality of parameter rules on the basis, and aiming at providing a learning method for pest tree sample identification: in order to obtain a rule with universality, the randomly selected pest tree sample is at least more than 1000; and carrying out digital or logical operation combination extraction on the wave spectrum values of different wave bands to obtain corresponding parameter rules, wherein the parameter selection method comprises a difference vegetation index, an RVI ratio vegetation index, a red-green ratio vegetation index and a normalized vegetation index. The above selection method is calculated as follows:
(a) Index of differential vegetation
DVI＝NIR-R…………………………………………………(1)
This index is extremely sensitive to changes in soil background. When vegetation is not completely covered, the soil background impact is greater. The vegetation index can be used for detecting the vegetation growth state, vegetation coverage, eliminating partial radiation errors and the like.
(b) RVI ratio vegetation index
RVI＝NIR/R…………………………………………………(2)
Generally, the RVI of green healthy vegetation covered areas is much greater than 1, while the RVI of non-vegetation covered ground (bare soil, man-made buildings, bodies of water, vegetation withered or severe insect pests) is around 1; RVI is a sensitive indicating parameter of green plants, has high correlation with LAI, leaf dry biomass (DM) and chlorophyll content, and can be used for detecting and estimating plant biomass; vegetation coverage affects RVI, which is very sensitive to vegetation when vegetation coverage is high, and which is significantly reduced when vegetation coverage is < 50%; RVI is affected by atmospheric conditions, atmospheric effects greatly reduce the sensitivity of the plant to be inspected, and atmospheric corrections, or calculations of RVI from reflectance, are required prior to calculation.
(c) Vegetation index of red-green ratio
RG＝R/G…………………………………………………(3)
Generally, the color of an object seen by the human eye is determined by the light it reflects. Most plants absorb red light and blue light and reflect green light, so most plants seen by people are green. The light reflected by plants with different growth conditions can also be different. Thus, the band ratio can also be used as a recognition factor for pest trees.
(d) Normalized vegetation index
NDVI＝(NIR-R)/(NIR+R)………………………(4)
The value of NDVI is defined between-1, negative values indicating a high reflection of visible light for ground coverage of clouds, water, snow, etc.; 0 represents rock or bare earth, etc., and NIR and R are approximately equal; positive values indicate vegetation coverage and increase with increasing coverage. Differences among crops with different growth conditions can be reflected through the NDVI value, and healthy and plant diseases and insect pests can be distinguished.
And finally, performing ground feature color feature extraction on all pixel points in the image sample to be classified by utilizing the training sample and the multiple parameter rules.
S2, screening out a target pixel set which accords with the characteristics of the multi-level diseased pine from the image sample to be classified after the ground object color characteristics are extracted; including the land features of target area, bare soil, water body, road, building construction, green grass and wheat field, rock road, forest shadow, etc.
And as shown in fig. 3, the step S2 includes the following sub-steps:
s21, filtering out pixels with NDVI values ranging from 0.25 to 0.55 to serve as a target area set 1;
s22, continuously filtering out the pixels with RVI values within the range of 1.3-2.8 from the residual pixel set to serve as a target area set 2;
s23, filtering out the pixel with the DVI value within the range of 150-400 from the residual set sample in the last step to serve as a target area set 3;
and S24, continuously calculating the pixel with the RGVI value within the range of more than 0.9 based on the union of the three output sets as a final screening result, namely the target pixel which accords with the characteristics of the disease tree.
And S3, carrying out primary denoising on the target pixel set conforming to the characteristics of the multilevel disease tree, and removing misjudged single points or sporadic points from the target pixel set conforming to the characteristics of the multilevel disease tree.
As shown in fig. 4, the step S3 includes the following sub-steps:
s31, traversing all target pixel sets, and searching adjacent pixel sets within the K value from four directions, namely the upper direction, the lower direction, the left direction and the right direction; if the set result is not null, allocating the pixels in the same set to the same cluster; if the set is empty, namely the set is a single point, setting the cluster number as 0; wherein the initial K value is 25 to 30.
S32, setting a clustering rule and a discrimination function of the category for two or more adjacent pixels, and outputting a clustered cluster center; the discrimination functions of the clustering rules and the categories are respectively merging rules and range discrimination functions, and the range discrimination function definition method can use either KNN or maximum likelihood method; the merge rule is defined as follows: judging whether the elements in the two cluster sets have intersection or not, and if so, combining the elements into a new cluster; and (4) calculating the number of neighbors in the K value range of each cluster member (the default value of the K value is 10), and outputting the point with the highest density as a cluster center.
And S4, carrying out secondary denoising on the target pixel set conforming to the characteristics of the multi-stage disease trees according to the distribution characteristics of the disease trees, and removing large-area noise points of non-forest regions, such as non-target objects like the grassland. The specific method comprises the following steps:
based on the distribution characteristics of the target points, calculating the number of the aggregated classifications in a certain area range (such as 100 pixels); and if the calculated classification number is more than 10, the region is considered to be not in accordance with the distribution characteristics of the sick trees in the forest region, and the target sample points in the region are considered as noise points.
And S5, carrying out three-time denoising on the target pixel set conforming to the characteristics of the multilevel disease tree according to the forest region background, and avoiding the misjudgment condition similar to the forest region edge.
And amplifying the edges of the classified clusters by taking the clusters as centers, firstly searching whether the proportion of the normal green trees in the radius R range is more than 20%, if so, retaining the classified clusters, and otherwise, taking the sample pixels in the clusters as noise points. And secondly, searching whether the proportion of the bare soil and the water body in the radius R range is larger than 1%, if so, regarding the sample pixel in the cluster as a noise point, and generally setting the radius R value to be 10-15 pixels.
The cluster center generated after the clustering and denoising is output to the SHP file, a classification result is generated, and the position of the cluster center is identified in the image file, as shown in fig. 5.
In order to verify the effectiveness of the classification denoising method for the high-resolution image of the diseased and dead pine in identifying the diseased and dead pine, a low-altitude collected double-spectrum remote sensing image sample is randomly selected for experiment, as shown in fig. 1. The method provided by the invention is based on comparative experimental analysis of a field actual measurement sample, a mixed-symptom matrix is used for comparing a classification result with real information of the ground surface, example data is shown in figure 6, and precision evaluation is calculated by using a Kappa coefficient. The calculation formula is as follows:
wherein p is o Is the sum of the number of correctly classified samples of each class divided by the total number of samples, i.e., the overall classification accuracy. p is a radical of e Is calculated as follows
Where N refers to the total number of pixels, X, of the test specimen i+ Means the total number of pixels, X, in the ith row +i Refers to the total number of pixels in the ith column. The following results are obtained after calculation:
the value of the Kappa coefficient is 0.83, so that the classification and denoising method for the high-resolution image of the dead pine diseases successfully classifies the disease and insect pest trees, and is a better classification method.
The invention relates to a classification denoising method and a classification denoising system for high-resolution images of dead pine diseases, which are used for extracting corresponding ground object color features of all pixel points in an image sample to be classified through classification training of the ground object color features according to verified spectrum information of disease and insect pests trees; then screening out a target pixel set which accords with the characteristics of the multi-stage disease pine from the image sample to be classified after the ground object color characteristics are extracted; carrying out primary denoising on the target pixel set conforming to the characteristics of the multilevel disease tree, and removing misjudged single points or sporadic points from the target pixel set conforming to the characteristics of the multilevel disease tree; secondly, carrying out secondary denoising on the target pixel set conforming to the characteristics of the multilevel disease trees according to the distribution characteristics of the disease trees, and removing large-area noise points in non-forest regions; and finally, carrying out three-time denoising on the target pixel set conforming to the characteristics of the multistage diseased trees according to the forest region background, avoiding the occurrence of misjudgment conditions similar to the forest region edge, and finally generating a classification result of the diseased and dead pine states, thereby realizing high detection precision identification of different disease areas and different disease degrees on the high-resolution images of the diseased and dead pine, and having high processing speed and high stability.
It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.
The above mainly describes a method for classifying and denoising high-resolution images of diseased and dead pine trees, and a system for classifying and denoising high-resolution images of diseased and dead pine trees is described in detail below.
Fig. 7 is a block diagram illustrating an embodiment of a classification denoising system for high-resolution images of dead pine tree diseases according to an embodiment of the present invention.
As shown in fig. 7, a classification denoising system for high-resolution image of dead pine comprises:
the color feature classification extraction module is used for performing corresponding ground object color feature extraction on all pixel points in the image sample to be classified according to the classification training of ground object color features of the verified pest and disease tree spectrum information;
the pixel screening module is used for screening a target pixel set which accords with the characteristics of the multi-level diseased pine from the image sample to be classified after the ground object color characteristics are extracted;
the primary denoising module is used for carrying out primary denoising on the target pixel set conforming to the multi-level disease tree characteristics and removing misjudged single points or sporadic points from the target pixel set conforming to the multi-level disease tree characteristics;
the secondary denoising module is used for carrying out secondary denoising on the target pixel set conforming to the multi-stage disease tree characteristics according to the disease tree distribution characteristics and removing large-area noise points of non-forest regions;
and the third denoising module is used for carrying out third denoising on the target pixel set conforming to the characteristics of the multilevel disease trees according to the forest region background so as to avoid the misjudgment condition similar to the forest region edge.
As shown in fig. 8, the pixel screening module includes the following functional units:
the NDVI value filtering unit is used for filtering out the pixels with the NDVI value ranging from 0.25 to 0.55 as a target area set 1;
an RVI value filtering unit for continuously filtering out the pixels with the RVI value within the range of 1.3-2.8 from the residual pixel set as a target area set 2;
a DVI value filtering unit for filtering out the pixels with DVI value in the range of 150-400 from the residual set samples in the last step as a target area set 3;
and the RGVI value filtering unit is used for continuously calculating the pixel with the RGVI value within the range of more than 0.9 based on the union set of the three output sets as a final screening result, namely a target pixel which accords with the characteristics of the disease tree.
As shown in fig. 9, the primary denoising module includes the following sub-steps:
the cluster allocation unit is used for traversing all target pixel sets and searching adjacent pixel sets within the K value from four directions, namely the upper direction, the lower direction, the left direction and the right direction; if the set result is not null, allocating the pixels in the same set to the same cluster; if the set is empty, the set is regarded as a single point, and the cluster number is set to be 0;
and the clustering unit is used for setting a clustering rule and a discrimination function of the category for two or more adjacent pixels and outputting a clustered cluster center.
It can be clearly understood by those skilled in the art that, for convenience and simplicity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or illustrated in a certain embodiment.
Those of ordinary skill in the art will appreciate that the modules, elements, and/or method steps of the various embodiments described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, all or part of the flow of the method according to the embodiments of the present invention may also be implemented by a computer program, which may be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the method embodiments described above may be implemented. . Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U.S. disk, removable hard disk, magnetic diskette, optical disk, computer Memory, read-Only Memory (ROM), random Access Memory (RAM), electrical carrier wave signal, telecommunications signal, and software distribution medium, etc. It should be noted that the computer readable medium may contain content that is subject to appropriate increase or decrease as required by legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer readable media does not include electrical carrier signals and telecommunications signals as is required by legislation and patent practice.
The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (2)
1. A classification denoising method for high-resolution images of dead pine trees is characterized by comprising the following steps:
s1, performing corresponding ground object color feature extraction on all pixel points in an image sample to be classified according to classification training of ground object color features of verified pest tree spectrum information;
s2, screening out a target pixel set which accords with the characteristics of the multi-stage disease pine from the image sample to be classified after the ground object color characteristics are extracted;
the step S2 comprises the following substeps:
s21, filtering out pixels with NDVI values ranging from 0.25 to 0.55 to serve as a target area set 1;
s22, continuously filtering out the pixels with RVI values within the range of 1.3-2.8 from the residual pixel set to serve as a target area set 2;
s23, filtering out the pixel with the DVI value within the range of 150-400 from the residual set sample in the last step to serve as a target area set 3;
s24, continuously calculating pixel with RGVI value within the range of more than 0.9 based on the union set of the three output sets as a final screening result, namely a target pixel conforming to the characteristics of the disease tree;
s3, carrying out primary denoising on the target pixel set conforming to the characteristics of the multilevel disease tree, and removing misjudged single points or sporadic points from the target pixel set conforming to the characteristics of the multilevel disease tree;
the step S3 includes the following substeps:
s31, traversing all target pixel sets, and searching adjacent pixel sets within the K value from four directions, namely the upper direction, the lower direction, the left direction and the right direction; if the set result is not null, allocating the pixels in the same set to the same cluster; if the set is empty, namely the set is a single point, setting the cluster number as 0;
s32, setting a clustering rule and a discrimination function of the category for two or more adjacent pixels, and outputting a clustered cluster center;
s4, performing secondary denoising on the target pixel set conforming to the characteristics of the multilevel disease trees according to the distribution characteristics of the disease trees, and removing large-area noise points in the non-forest region;
and S5, carrying out three-time denoising on the target pixel set conforming to the characteristics of the multilevel disease trees according to the forest region background, and avoiding the misjudgment condition similar to the forest region edge.
2. A classification denoising system for high-resolution images of dead pine trees, comprising:
the color feature classification extraction module is used for performing corresponding ground object color feature extraction on all pixel points in the image sample to be classified according to the classification training of ground object color features of the verified pest and disease tree spectrum information;
the pixel screening module is used for screening a target pixel set which accords with the characteristics of the multi-level disease pine from the image sample to be classified after the ground object color characteristics are extracted;
the pixel screening module comprises the following functional units:
the NDVI value filtering unit is used for filtering out the pixels with the NDVI value ranging from 0.25 to 0.55 as a target area set 1;
the RVI value filtering unit is used for continuously filtering out the pixels with the RVI value within the range of 1.3-2.8 from the residual pixel set to be used as a target area set 2;
a DVI value filtering unit for filtering out the pixels with DVI value in the range of 150-400 from the residual set samples in the last step as a target area set 3;
the RGVI value filtering unit is used for continuously calculating pixel with RGVI value within the range of more than 0.9 based on the union set of the three output sets as a final screening result, namely a target pixel conforming to the characteristics of the disease tree;
the primary denoising module is used for carrying out primary denoising on the target pixel set conforming to the characteristics of the multilevel disease tree, and removing misjudged single points or sporadic points from the target pixel set conforming to the characteristics of the multilevel disease tree;
the primary denoising module comprises the following substeps:
the cluster distribution unit is used for traversing all target pixel sets and searching adjacent pixel sets within the K value from four directions, namely upper, lower, left and right directions; if the set result is not null, allocating the pixels in the same set to the same cluster; if the set is empty, namely the set is a single point, setting the cluster number as 0;
the clustering unit is used for setting a clustering rule and a discrimination function of a category for two or more adjacent pixels and outputting a clustered cluster center;
the secondary denoising module is used for carrying out secondary denoising on the target pixel set conforming to the characteristics of the multilevel disease trees according to the distribution characteristics of the disease trees, and removing large-area noise points in the non-forest region;
and the third denoising module is used for carrying out three times of denoising on the target pixel set conforming to the characteristics of the multilevel disease trees according to the forest region background so as to avoid the misjudgment condition similar to the forest region edge.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810364964.2A CN108764284B (en) | 2018-04-23 | 2018-04-23 | Classification and denoising method and system for high-resolution image of dead pine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810364964.2A CN108764284B (en) | 2018-04-23 | 2018-04-23 | Classification and denoising method and system for high-resolution image of dead pine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108764284A CN108764284A (en) | 2018-11-06 |
| CN108764284B true CN108764284B (en) | 2022-11-22 |
Family
ID=64011040
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810364964.2A Active CN108764284B (en) | 2018-04-23 | 2018-04-23 | Classification and denoising method and system for high-resolution image of dead pine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108764284B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110096970A (en) * | 2019-04-12 | 2019-08-06 | 苏州市职业大学 | Pine forest discoloration standing tree single plant identification method based on WV3 satellite image |
| CN110533595A (en) * | 2019-09-03 | 2019-12-03 | 北京麦飞科技有限公司 | A kind of method and system of crop disease and insect Spatial distributions distribution monitoring |
| CN111582176A (en) * | 2020-05-09 | 2020-08-25 | 湖北同诚通用航空有限公司 | Visible light remote sensing image withered and dead wood recognition software system and recognition method |
| CN113197578B (en) * | 2021-05-07 | 2023-06-09 | 天津医科大学 | Schizophrenia classification method and system based on multi-center model |
| CN113609913B (en) * | 2021-07-08 | 2022-08-19 | 三峡大学 | Pine wood nematode disease tree detection method based on sampling threshold interval weighting |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101770577A (en) * | 2010-01-18 | 2010-07-07 | 浙江林学院 | Method for extracting information on expansion and elimination of dead wood of pine wilt disease in air photos of unmanned aerial vehicle |
| CN102937574A (en) * | 2012-07-20 | 2013-02-20 | 北京农业信息技术研究中心 | Information extraction method for plant diseases and insect pests based on satellite images |
| CN106915462A (en) * | 2017-02-14 | 2017-07-04 | 福建兴宇信息科技有限公司 | Forestry pests ＆ diseases intelligent identifying system based on multi-source image information |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8639639B2 (en) * | 2009-08-31 | 2014-01-28 | Bhagwan Mahavir Medical Research Centre | Predicting possible outcomes in multi-factored diseases |
-
2018
- 2018-04-23 CN CN201810364964.2A patent/CN108764284B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101770577A (en) * | 2010-01-18 | 2010-07-07 | 浙江林学院 | Method for extracting information on expansion and elimination of dead wood of pine wilt disease in air photos of unmanned aerial vehicle |
| CN102937574A (en) * | 2012-07-20 | 2013-02-20 | 北京农业信息技术研究中心 | Information extraction method for plant diseases and insect pests based on satellite images |
| CN106915462A (en) * | 2017-02-14 | 2017-07-04 | 福建兴宇信息科技有限公司 | Forestry pests ＆ diseases intelligent identifying system based on multi-source image information |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108764284A (en) | 2018-11-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108764284B (en) | Classification and denoising method and system for high-resolution image of dead pine | |
| Yuan et al. | Detection of anthracnose in tea plants based on hyperspectral imaging | |
| Puletti et al. | Unsupervised classification of very high remotely sensed images for grapevine rows detection | |
| US6282301B1 (en) | Ares method of sub-pixel target detection | |
| CN112101159B (en) | Multi-temporal forest remote sensing image change monitoring method | |
| Zhang et al. | Robust hyperspectral vision-based classification for multi-season weed mapping | |
| CN111091052A (en) | Corn lodging area extraction system and method based on maximum likelihood method | |
| WO2001033505A2 (en) | Multi-variable model for identifying crop response zones in a field | |
| CN110533595A (en) | A kind of method and system of crop disease and insect Spatial distributions distribution monitoring | |
| CN112183209A (en) | Regional crop classification method and system based on multi-dimensional feature fusion | |
| CN108710864B (en) | Winter wheat remote sensing extraction method based on multi-dimensional identification and image noise reduction processing | |
| CN106373150A (en) | Identification method of corns for seed on the basis of high resolution remote sensing data texture analysis | |
| CN105067532A (en) | Method for identifying early-stage disease spots of sclerotinia sclerotiorum and botrytis of rape | |
| CN112766036A (en) | Remote sensing extraction method and device for lodging corn | |
| CN116543316B (en) | Method for identifying turf in paddy field by utilizing multi-time-phase high-resolution satellite image | |
| CN112861810A (en) | Artificial forest planting time automatic detection method based on time sequence remote sensing observation data | |
| CN114581768A (en) | Crop lodging unmanned aerial vehicle monitoring method and device | |
| CN108764285A (en) | A kind of recognition methods of pine tree Deceased wood and system based on high resolution image | |
| JP3564540B2 (en) | Land use classification processing method using SAR image | |
| CN116310864A (en) | Automatic identification method, system, electronic equipment and medium for crop lodging | |
| CN109993228A (en) | Plant protection drone rice sheath blight disease recognition methods based on machine vision | |
| CN115631419A (en) | Method and device for extracting rice planting area and spatial distribution based on change detection | |
| CN110096970A (en) | Pine forest discoloration standing tree single plant identification method based on WV3 satellite image | |
| CN109948556B (en) | Crop classification identification method and system | |
| Raghavendra | Multivariant disease detection from different plant leaves and classification using multiclass support vector machine |
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 |