CN117935054A - Rice identification method, system and device based on novel red-edge rice index - Google Patents
Rice identification method, system and device based on novel red-edge rice index Download PDFInfo
- Publication number
- CN117935054A CN117935054A CN202410108620.0A CN202410108620A CN117935054A CN 117935054 A CN117935054 A CN 117935054A CN 202410108620 A CN202410108620 A CN 202410108620A CN 117935054 A CN117935054 A CN 117935054A
- Authority
- CN
- China
- Prior art keywords
- rice
- red
- index
- edge
- paddy
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 235000007164 Oryza sativa Nutrition 0.000 title claims abstract description 318
- 241000209094 Oryza Species 0.000 title claims abstract description 292
- 235000009566 rice Nutrition 0.000 title claims abstract description 292
- 238000000034 method Methods 0.000 title claims abstract description 30
- 230000011218 segmentation Effects 0.000 claims abstract description 53
- 238000013145 classification model Methods 0.000 claims abstract description 30
- 235000007189 Oryza longistaminata Nutrition 0.000 claims abstract description 26
- 240000007594 Oryza sativa Species 0.000 claims abstract description 26
- 238000012549 training Methods 0.000 claims abstract description 25
- 238000009826 distribution Methods 0.000 claims abstract description 21
- 238000012216 screening Methods 0.000 claims abstract description 14
- 238000004590 computer program Methods 0.000 claims description 13
- 230000003595 spectral effect Effects 0.000 claims description 9
- 238000007637 random forest analysis Methods 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 7
- 238000003860 storage Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 abstract description 4
- 230000035945 sensitivity Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 10
- 238000003973 irrigation Methods 0.000 description 5
- 230000002262 irrigation Effects 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 238000012795 verification Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000009331 sowing Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 244000037666 field crops Species 0.000 description 2
- 238000003306 harvesting Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 238000007792 addition Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000009355 double cropping Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Image Processing (AREA)
Abstract
The invention discloses a rice identification method, a system and a device based on a novel red-edge rice index, wherein the method comprises the following steps: acquiring first rice image data, screening, and classifying rice samples to obtain potential rice areas and non-rice areas; obtaining second rice image data to obtain a plurality of independent segmentation objects; combining the independent segmentation object with the potential rice area to obtain a potential rice area distribution range; constructing a red rice index model to obtain a red rice index image, and optimizing the red rice index image to obtain an optimized red rice index image; and training based on the optimized red-edge paddy index image and image data of paddy fields of different types to obtain a paddy field classification model, and obtaining the types of paddy fields in the image of the paddy field to be detected based on the paddy field classification model. The method can solve the supersaturation phenomenon existing in the traditional normalized vegetation index, enhance the sensitivity of the dense areas of the vegetation index, and finally improve the extraction precision of rice types.
Description
Technical Field
The invention relates to the technical field of big data processing, in particular to a novel red-edge rice index-based rice identification method, system and device.
Background
Along with the development of technology, the remote sensing satellite technology has become a main method for extracting rice areas, and vegetation indexes integrate information of a plurality of spectral bands, so that a large amount of work has been carried out on the basis of the vegetation index-based rice identification method. At present, a remote sensing image with high discrimination degree is selected according to the rice weather characteristics, and a planting area of rice is extracted according to a vegetation index in a corresponding period.
The presently commonly used normalization indexes, such as NDVI, GNDVI and NDRE, are very sensitive to the perception of green vegetation, and therefore, a supersaturation phenomenon occurs, which leads to reduced sensitivity in dense areas of vegetation, and also leads to overestimation of vegetation coverage in the early and final stages of growing seasons, which seriously affects the accuracy of subsequent identification.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a novel red-edge rice index-based rice identification method, system and device.
In order to solve the technical problems, the invention is solved by the following technical scheme:
a novel red-edge rice index-based rice identification method comprises the following steps:
Acquiring first rice image data, screening, and taking the image data lower than a preset scattering coefficient threshold value as a rice sample;
classifying the rice samples to obtain potential rice areas and non-rice areas;
acquiring second rice image data and performing object-oriented segmentation to obtain a plurality of independent segmentation objects;
combining the independent segmentation object with the potential rice area to obtain a potential rice area distribution range;
constructing a red rice index model, obtaining a red rice index image of the image data of the rice to be detected based on the red rice index model, and optimizing the red rice index image based on the distribution range of the potential rice area to obtain an optimized red rice index image;
and training based on the optimized red-edge paddy index image and image data of paddy fields of different types to obtain a paddy field classification model, and obtaining the types of paddy fields in the image of the paddy field to be detected based on the paddy field classification model.
As an implementation manner, the rice sample set is classified based on a random forest algorithm to obtain a classified rice sample, a specific pixel value is set for each pixel in the classified rice sample, the pixel value is 1 to represent a potential rice area, and the pixel value is 0 to represent a non-rice area.
As an implementation manner, the object-oriented segmentation is performed to obtain a plurality of independent segmented objects, which includes the following steps:
Presetting optimal segmentation parameters, and segmenting the second rice image data based on the optimal segmentation parameters to obtain a plurality of initial independent segmentation objects;
and classifying the plurality of initial independent segmentation objects based on a preset classification rule to obtain independent segmentation objects.
As an implementation manner, the classification rule includes:
Counting the pixel number of each initial independent segmentation object to obtain a pixel counting result;
if the pixel statistics result is not smaller than the preset pixel threshold value, the potential rice area is marked.
As an embodiment, the construction of the red-edge rice index model includes the following steps:
analyzing spectral characteristics of rice, and selecting red wave bands, near infrared wave bands and red edge wave bands;
constructing a red-edge rice index model based on a red wave band, a near infrared wave band and a red-edge wave band, wherein the red-edge rice index model is as follows:
wherein ρ nir,ρred-edge2,ρred represents the near infrared band, the red side band and the red band, respectively.
As an implementation manner, the method for obtaining the rice classification model based on the optimized red-edge rice index image and the image data training of different paddy fields, and obtaining the types of the rice in the image of the paddy field to be detected based on the rice classification model comprises the following steps:
acquiring image data of paddy fields of different types;
Constructing a rice classification pre-training model;
training the rice classification pre-training model based on the image data of the paddy fields of different types and the optimized red-edge rice index image to obtain a rice classification model;
And identifying the types of the rice in the rice field image to be detected based on the rice classification model to obtain the types of the rice in the rice field image to be detected.
A novel red-edge rice index-based rice identification system comprises an acquisition screening module, a first classification module, an acquisition segmentation module, a data combination module, a model construction module and a training identification module;
The acquisition and screening module is used for acquiring first rice image data and screening, and taking the image data lower than a preset scattering coefficient threshold value as a rice sample;
the first classification module is used for classifying the rice samples to obtain potential rice areas and non-rice areas;
the acquisition and segmentation module is used for acquiring second rice image data and performing object-oriented segmentation to obtain a plurality of independent segmentation objects;
The data combination module is used for combining the independent segmentation object with the potential rice area to obtain a potential rice area distribution range;
the model construction module is used for constructing a red-edge rice index model, obtaining red-edge rice index images of the rice image data to be detected based on the red-edge rice index model, and optimizing the red-edge rice index images based on the distribution range of the potential rice areas to obtain optimized red-edge rice index images;
The training recognition module trains based on the optimized red-edge paddy index image and the image data of paddy fields of different types to obtain a paddy field classification model, and obtains the types of paddy fields in the image of the paddy field to be detected based on the paddy field classification model.
As an embodiment, the model building module is configured to:
analyzing spectral characteristics of rice, and selecting red wave bands, near infrared wave bands and red edge wave bands;
constructing a red-edge rice index model based on a red wave band, a near infrared wave band and a red-edge wave band, wherein the red-edge rice index model is as follows:
wherein ρ nir,ρred-edge2,ρred represents the near infrared band, the red side band and the red band, respectively.
A computer readable storage medium storing a computer program which when executed by a processor performs the method of:
Acquiring first rice image data, screening, and taking the image data lower than a preset scattering coefficient threshold value as a rice sample;
classifying the rice samples to obtain potential rice areas and non-rice areas;
acquiring second rice image data and performing object-oriented segmentation to obtain a plurality of independent segmentation objects;
combining the independent segmentation object with the potential rice area to obtain a potential rice area distribution range;
constructing a red rice index model, obtaining a red rice index image of the image data of the rice to be detected based on the red rice index model, and optimizing the red rice index image based on the distribution range of the potential rice area to obtain an optimized red rice index image;
and training based on the optimized red-edge paddy index image and image data of paddy fields of different types to obtain a paddy field classification model, and obtaining the types of paddy fields in the image of the paddy field to be detected based on the paddy field classification model.
A novel red-edge rice index-based rice identification device, comprising a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor implements the following method when executing the computer program:
Acquiring first rice image data, screening, and taking the image data lower than a preset scattering coefficient threshold value as a rice sample;
classifying the rice samples to obtain potential rice areas and non-rice areas;
acquiring second rice image data and performing object-oriented segmentation to obtain a plurality of independent segmentation objects;
combining the independent segmentation object with the potential rice area to obtain a potential rice area distribution range;
constructing a red rice index model, obtaining a red rice index image of the image data of the rice to be detected based on the red rice index model, and optimizing the red rice index image based on the distribution range of the potential rice area to obtain an optimized red rice index image;
and training based on the optimized red-edge paddy index image and image data of paddy fields of different types to obtain a paddy field classification model, and obtaining the types of paddy fields in the image of the paddy field to be detected based on the paddy field classification model.
The invention has the remarkable technical effects due to the adoption of the technical scheme:
The method can solve the supersaturation phenomenon existing in the traditional normalized vegetation index, enhance the sensitivity of the dense areas of the vegetation index, and finally improve the extraction precision of rice types.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, and that other drawings can be obtained according to these drawings without inventive faculty for a person skilled in the art.
FIG. 1 is a schematic overall flow diagram of the present invention;
FIG. 2 is a schematic flow chart of the method after combining test data;
FIG. 3 is a schematic view of spectral features of rice;
FIG. 4 is a schematic diagram of the rice identification result according to the present invention;
FIG. 5 is a schematic diagram of image contrast according to the present invention;
fig. 6 is a schematic diagram of the overall structure of the system of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the following examples, which are illustrative of the present invention and are not intended to limit the present invention thereto.
Example 1:
A novel red-edge rice index-based rice identification method is shown in fig. 1, and comprises the following steps:
S100, acquiring first rice image data, screening, and taking the image data lower than a preset scattering coefficient threshold value as a rice sample;
S200, classifying the rice samples to obtain potential rice areas and non-rice areas;
s300, acquiring second rice image data and performing object-oriented segmentation to obtain a plurality of independent segmentation objects;
s400, combining the independent segmentation object with the potential rice area to obtain a potential rice area distribution range;
S500, constructing a red-edge rice index model, obtaining a red-edge rice index image of the rice image data to be measured based on the red-edge rice index model, and optimizing the red-edge rice index image based on a potential rice area distribution range to obtain an optimized red-edge rice index image;
s600, training based on the optimized red-edge paddy index image and image data of paddy fields of different types to obtain a paddy field classification model, and obtaining the types of paddy fields in the image of the paddy field to be detected based on the paddy field classification model.
In one embodiment, a random forest algorithm is adopted to classify the rice sample set to obtain classified rice samples, a specific pixel value is set for each pixel in the classified rice samples, the pixel value is 1 to represent a potential rice area, and the pixel value is 0 to represent a non-rice area. In the setting of the parameters of the random forest algorithm, the number of trees in the random forest algorithm is set to 100, and the rest parameters are default, so that the setting can be verified for many times to realize higher precision with higher efficiency. In the actual test process, visual interpretation is carried out according to the features of the paddy irrigation period based on the images of the first number of the multi-timing sentry, potential paddy and non-paddy samples are checked, and the extraction of potential paddy areas is preliminarily realized based on a random forest classification algorithm. Since paddy fields contain a large amount of water in the irrigation period, the backscattering coefficient value in the period is lower than that in other growth periods, and therefore paddy fields can be distinguished from dry crops according to the characteristics, and the irrigation signal is a phenomenon that when the paddy fields are submerged in water in the irrigation period, the backscattering coefficient value is lower than that in other growth periods. This is a unique feature of paddy field crops, and the back scattering coefficient value of dry field crops will not have large fluctuation. Based on the long-time-sequence sentinel first-number image, pixels with the backward scattering coefficient value obviously lower than that of other periods in the partial irrigation period are selected as potential rice samples, and otherwise, the potential rice samples are not potential rice samples. Then, random forest classification is carried out, each pixel in the classified image has a unique pixel value, wherein a pixel value of 1 is a potential rice area, and a pixel value of 0 is a non-rice area. Thus, the potential rice area distribution range is obtained initially.
In step S300, the object-oriented segmentation is performed to obtain a plurality of independent segmented objects, which includes the following steps:
Presetting optimal segmentation parameters, and segmenting the second rice image data based on the optimal segmentation parameters to obtain a plurality of initial independent segmentation objects;
and classifying the plurality of initial independent segmentation objects based on a preset classification rule to obtain independent segmentation objects.
And (3) performing object-oriented segmentation on the sentinel second image by using Yikang software to generate a plurality of independent objects. And combining the initially extracted potential rice area results, and optimizing the potential rice area to obtain a final potential rice area range. The method comprises the steps of performing object-oriented segmentation on a sentinel second image by using Yikang software, comparing results of different segmentation parameters (shape, scale, compactness) with a paddy field, and finally selecting parameters (scale is 10, shape is 0.1 and compatibility is 0.6) which are most suitable for the area, wherein the parameter combination can realize complete presentation of the paddy field. The segmented independent objects are then combined with the primary extracted potential rice field results, that is, the potential rice field results are two, one is potential rice and the other is non-rice, corresponding to 1 and 0 respectively. We count the number of pixel values of 1 and 0 in each independent object.
The classification rule includes: counting the pixel number of each initial independent segmentation object to obtain a pixel counting result; if the pixel statistics result is not smaller than the preset pixel threshold value, the potential rice area is marked. The method comprises the following steps: and counting the number of pixels in each independent land, wherein the ratio of which pixel value is larger than 50%, and which pixel value is the label value of the land. (if both pixel values in a plot account for 50%, they are also demarcated as potential rice areas). The optimization of the potential rice area is realized based on the rule, and the final potential rice area distribution range is obtained.
In step S500, the building of the red-edge rice index model includes the following steps:
Any rice image data is obtained, and of course, the first rice image data and the second rice image data can also be obtained, the spectral characteristics of rice are analyzed, and red wave bands, near infrared wave bands and red edge wave bands are selected, as shown in fig. 3;
constructing a red-edge rice index model based on a red wave band, a near infrared wave band and a red-edge wave band, wherein the red-edge rice index model is as follows:
wherein ρ nir,ρred-edge2,ρred represents the near infrared band, the red side band and the red band, respectively.
Through analysis of rice spectral characteristics, red wave bands, near infrared wave bands and red edge wave bands are selected to construct red edge rice indexes. According to the growth characteristics of single-season rice and double-season rice (sowing early-season rice in 4 months, sowing early-season rice in middle ten days, transplanting early 5 months, harvesting late-season rice in 7 months, transplanting late-season rice immediately (double robbing), generally, late-season rice must be harvested gradually from 10 months to 11 months before autumn, sowing early-season rice in front of and after Qing, transplanting early-time five months and harvesting early-season rice in 8 months), and combining available remote sensing images, we select high-resolution six images of 19 days and 5 days of 8 months, and sentinel two images of 29 days and 14 days of 9 months and 10 months, and calculate red-edge rice indexes of the images in 4 periods.
In step S600, the training is performed based on the optimized red-edge paddy index image and the image data of paddy fields of different types to obtain a paddy classification model, and the types of paddy in the image of the paddy field to be measured are obtained based on the paddy classification model, which comprises the following steps:
acquiring image data of paddy fields of different types;
Constructing a rice classification pre-training model;
training the rice classification pre-training model based on the image data of the paddy fields of different types and the optimized red-edge rice index image to obtain a rice classification model;
And identifying the types of the rice in the rice field image to be detected based on the rice classification model to obtain the types of the rice in the rice field image to be detected.
The four original remote sensing images and the optimized red-edge rice index images are used as machine learning characteristics, and a random forest classification algorithm is used for training a prediction model. And then, the model is used for carrying out prediction classification on the research area to preliminarily obtain the distribution ranges of the single-cropping rice, the double-cropping rice and the non-rice. Finally, the result is subjected to conditional selection: 1) When a pixel is non-rice in the result of the potential rice area, the pixel is non-rice; 2) When a pixel is potential rice in the result of the potential rice field, the value of the pixel depends on the predicted value of the model. Based on this conditional selection method, the final single-, double-and non-rice distribution ranges were obtained, as shown in fig. 4.
Finally, the results of the rice identification are also verified with respect to accuracy using the overall accuracy, kappa coefficient, producer accuracy, consumer accuracy and F1 score. In combination with the actual case, the overall flow chart is shown in fig. 2.
The invention mainly uses the precision of the rice identification result to verify the validity of the red-edge rice index. Specifically, red-edge rice indexes and common normalized vegetation indexes are used for extracting the range of a rice planting area, and then accuracy verification indexes are used for comparing the accuracy of rice planting area results extracted by different indexes. As shown in table 1:
table 1NDVI, GNDVI, NDRE, and RERI verification of accuracy of Rice identification results
Based on the results of the red-edge rice index, all accuracy verification indexes are higher than other indexes. Through the verification of the accuracy of the result, the red-edge rice index provided by us can be obtained to have certain effectiveness in the extraction of the rice planting area.
In addition, the invention also verifies the advantage of the red-edge rice index in overcoming the supersaturation phenomenon of the normalized vegetation index by comparing the images of the indexes.
In fig. 5, it is worth explaining that the higher the index value, the more "highlighting" effect is shown in the image contrast chart, and it can be seen that the index RERI provided by the present invention can well distinguish rice from other green vegetation. The concrete implementation is as follows: the position of the rice is highlighted, while images of other indexes show supersaturation, the rice and other crops are highlighted, or the area of the rice which is highlighted is few, so that the rice and other crops are difficult to be roughly distinguished by naked eyes. It can be seen that the red rice index calculated by the red rice index model created by the invention has better characteristics than other indexes in the accuracy verification of the rice identification result or the display of images, thereby verifying the effectiveness of the invention.
Example 2:
A novel red-edge rice index-based rice identification system is shown in FIG. 6, and comprises an acquisition screening module 100, a first classification module 200, an acquisition segmentation module 300, a data combination module 400, a model construction module 500 and a training identification module 600;
the acquiring and screening module 100 is configured to acquire and screen first rice image data, and take the image data lower than a preset scattering coefficient threshold value as a rice sample;
the first classification module 200 is configured to classify the rice sample to obtain a potential rice area and a non-rice area;
The obtaining and dividing module 300 is configured to obtain second rice image data and perform object-oriented division to obtain a plurality of independent division objects;
the data combination module 400 is configured to combine the independent segmentation object with the potential rice area to obtain a distribution range of the potential rice area;
The model construction module 500 is configured to construct a red-edge rice index model, obtain a red-edge rice index image of the image data of the rice to be measured based on the red-edge rice index model, and optimize the red-edge rice index image based on the distribution range of the potential rice area to obtain an optimized red-edge rice index image;
The training recognition module 600 trains to obtain a rice classification model based on the optimized red-edge rice index image and the image data of different types of paddy fields, and obtains the types of rice in the image of the paddy field to be detected based on the rice classification model.
In one embodiment, the model building module 500 is configured to:
analyzing spectral characteristics of rice, and selecting red wave bands, near infrared wave bands and red edge wave bands;
constructing a red-edge rice index model based on a red wave band, a near infrared wave band and a red-edge wave band, wherein the red-edge rice index model is as follows:
wherein ρ nir,ρred-edge2,ρred represents the near infrared band, the red side band and the red band, respectively.
All changes and modifications that come within the spirit and scope of the invention are desired to be protected and all equivalent thereto are deemed to be within the scope of the invention.
In this specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different manner from other embodiments, so that identical and similar parts of each embodiment are mutually referred to.
It will be apparent to those skilled in the art that embodiments of the present invention may be provided as a method, apparatus, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, terminal devices (systems), and computer program products according to the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It should be noted that:
Reference in the specification to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the invention. Thus, the appearances of the phrase "one embodiment" or "an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment.
In addition, the specific embodiments described in the present specification may differ in terms of parts, shapes of components, names, and the like. All equivalent or simple changes of the structure, characteristics and principle according to the inventive concept are included in the protection scope of the present invention. Those skilled in the art may make various modifications or additions to the described embodiments or substitutions in a similar manner without departing from the scope of the invention as defined in the accompanying claims.
Claims (10)
1. The novel red-edge rice index-based rice identification method is characterized by comprising the following steps of:
Acquiring first rice image data, screening, and taking the image data lower than a preset scattering coefficient threshold value as a rice sample;
classifying the rice samples to obtain potential rice areas and non-rice areas;
acquiring second rice image data and performing object-oriented segmentation to obtain a plurality of independent segmentation objects;
combining the independent segmentation object with the potential rice area to obtain a potential rice area distribution range;
constructing a red rice index model, obtaining a red rice index image of the image data of the rice to be detected based on the red rice index model, and optimizing the red rice index image based on the distribution range of the potential rice area to obtain an optimized red rice index image;
and training based on the optimized red-edge paddy index image and image data of paddy fields of different types to obtain a paddy field classification model, and obtaining the types of paddy fields in the image of the paddy field to be detected based on the paddy field classification model.
2. The method for identifying rice based on the novel red-edge rice index according to claim 1, wherein the rice sample set is classified based on a random forest algorithm to obtain a classified rice sample, a specific pixel value is set for each pixel in the classified rice sample, wherein a pixel value of 1 represents a potential rice area, and a pixel value of 0 represents a non-rice area.
3. The method for identifying rice based on the novel red-edge rice index according to claim 1, wherein said object-oriented segmentation is performed to obtain a plurality of independent segmented objects, comprising the steps of:
Presetting optimal segmentation parameters, and segmenting the second rice image data based on the optimal segmentation parameters to obtain a plurality of initial independent segmentation objects;
and classifying the plurality of initial independent segmentation objects based on a preset classification rule to obtain independent segmentation objects.
4. A novel red-edged rice index-based rice identification method according to claim 3, wherein the classification rules include:
Counting the pixel number of each initial independent segmentation object to obtain a pixel counting result;
if the pixel statistics result is not smaller than the preset pixel threshold value, the potential rice area is marked.
5. The method for identifying rice based on a novel red-edge rice index according to claim 1, wherein the constructing of the red-edge rice index model comprises the steps of:
analyzing spectral characteristics of rice, and selecting red wave bands, near infrared wave bands and red edge wave bands;
constructing a red-edge rice index model based on a red wave band, a near infrared wave band and a red-edge wave band, wherein the red-edge rice index model is as follows:
Wherein ρ nir,ρred-edge2,ρred represents the near infrared band, the red side band and the red band, respectively.
6. The method for identifying paddy rice based on the novel red edge paddy rice index according to claim 1, wherein the image data training based on the optimized red edge paddy rice index image and paddy fields of different types is used for obtaining a paddy rice classification model, and the type of paddy rice in the image of the paddy field to be detected is obtained based on the paddy rice classification model, comprising the following steps:
acquiring image data of paddy fields of different types;
Constructing a rice classification pre-training model;
training the rice classification pre-training model based on the image data of the paddy fields of different types and the optimized red-edge rice index image to obtain a rice classification model;
And identifying the types of the rice in the rice field image to be detected based on the rice classification model to obtain the types of the rice in the rice field image to be detected.
7. The novel red-edge rice index-based rice identification system is characterized by comprising an acquisition screening module, a first classification module, an acquisition segmentation module, a data combination module, a model construction module and a training identification module;
The acquisition and screening module is used for acquiring first rice image data and screening, and taking the image data lower than a preset scattering coefficient threshold value as a rice sample;
the first classification module is used for classifying the rice samples to obtain potential rice areas and non-rice areas;
the acquisition and segmentation module is used for acquiring second rice image data and performing object-oriented segmentation to obtain a plurality of independent segmentation objects;
The data combination module is used for combining the independent segmentation object with the potential rice area to obtain a potential rice area distribution range;
the model construction module is used for constructing a red-edge rice index model, obtaining red-edge rice index images of the rice image data to be detected based on the red-edge rice index model, and optimizing the red-edge rice index images based on the distribution range of the potential rice areas to obtain optimized red-edge rice index images;
The training recognition module trains based on the optimized red-edge paddy index image and the image data of paddy fields of different types to obtain a paddy field classification model, and obtains the types of paddy fields in the image of the paddy field to be detected based on the paddy field classification model.
8. The novel red-edged rice index-based rice identification system of claim 7, wherein the model building module is configured to:
analyzing spectral characteristics of rice, and selecting red wave bands, near infrared wave bands and red edge wave bands;
constructing a red-edge rice index model based on a red wave band, a near infrared wave band and a red-edge wave band, wherein the red-edge rice index model is as follows:
Wherein ρ nir,ρred-edge2,ρred represents the near infrared band, the red side band and the red band, respectively.
9. A computer readable storage medium storing a computer program, which when executed by a processor implements the method of any one of claims 1 to 6.
10. A novel red-edged rice index-based rice identification device comprising a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor implements the method according to any one of claims 1 to 6 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410108620.0A CN117935054A (en) | 2024-01-25 | 2024-01-25 | Rice identification method, system and device based on novel red-edge rice index |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410108620.0A CN117935054A (en) | 2024-01-25 | 2024-01-25 | Rice identification method, system and device based on novel red-edge rice index |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117935054A true CN117935054A (en) | 2024-04-26 |
Family
ID=90760808
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410108620.0A Pending CN117935054A (en) | 2024-01-25 | 2024-01-25 | Rice identification method, system and device based on novel red-edge rice index |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117935054A (en) |
-
2024
- 2024-01-25 CN CN202410108620.0A patent/CN117935054A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112101159B (en) | Multi-temporal forest remote sensing image change monitoring method | |
| Liu et al. | A vision-based robust grape berry counting algorithm for fast calibration-free bunch weight estimation in the field | |
| CN107330413B (en) | A kind of mother drug plants recognition methods based on remote sensing technology | |
| Zheng et al. | Research on tomato detection in natural environment based on RC-YOLOv4 | |
| CN108710864B (en) | Winter wheat remote sensing extraction method based on multi-dimensional identification and image noise reduction processing | |
| CN112699756B (en) | Hyperspectral image-based tea origin identification method and system | |
| CN108764284B (en) | Classification and denoising method and system for high-resolution image of dead pine | |
| CN116543316B (en) | Method for identifying turf in paddy field by utilizing multi-time-phase high-resolution satellite image | |
| CN111918547B (en) | Crown recognition device, crown recognition method, program, and recording medium | |
| Palacios et al. | Early yield prediction in different grapevine varieties using computer vision and machine learning | |
| CN113963260A (en) | Extraction method and device for winter wheat planting area and computer equipment | |
| CN111007013A (en) | Crop rotation fallow remote sensing monitoring method and device for northeast cold region | |
| JP2008152425A (en) | Agricultural land lot data production system | |
| CN116129260A (en) | Forage grass image recognition method based on deep learning | |
| CN109034189A (en) | Forest type identification method based on high score remote sensing image | |
| Sabzchi-Dehkharghani et al. | Recognition of different yield potentials among rain-fed wheat fields before harvest using remote sensing | |
| Kamal et al. | A preliminary study on machine learning and google earth engine for mangrove mapping | |
| CN107103295A (en) | Optical remote sensing image cloud detection method of optic | |
| CN117589767A (en) | Tobacco leaf harvesting time determining method, electronic equipment and storage medium | |
| CN110059890B (en) | County scale agricultural flood monitoring method and system | |
| CN112528726A (en) | Aphis gossypii insect pest monitoring method and system based on spectral imaging and deep learning | |
| CN117935054A (en) | Rice identification method, system and device based on novel red-edge rice index | |
| Moradi et al. | Potential evaluation of visible-thermal UAV image fusion for individual tree detection based on convolutional neural network | |
| CN109784168A (en) | A kind of high-definition remote sensing passway for transmitting electricity inspection method and system | |
| Huang et al. | Recognition and counting of pitaya trees in karst mountain environment based on unmanned aerial vehicle RGB images |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination |