CN115349340B - Sorghum fertilization control method and system based on artificial intelligence - Google Patents
Sorghum fertilization control method and system based on artificial intelligence Download PDFInfo
- Publication number
- CN115349340B CN115349340B CN202211138982.1A CN202211138982A CN115349340B CN 115349340 B CN115349340 B CN 115349340B CN 202211138982 A CN202211138982 A CN 202211138982A CN 115349340 B CN115349340 B CN 115349340B
- Authority
- CN
- China
- Prior art keywords
- area
- image
- sorghum
- ndvi
- amount
- 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
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01C—PLANTING; SOWING; FERTILISING
- A01C23/00—Distributing devices specially adapted for liquid manure or other fertilising liquid, including ammonia, e.g. transport tanks or sprinkling wagons
- A01C23/007—Metering or regulating systems
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01M—CATCHING, TRAPPING OR SCARING OF ANIMALS; APPARATUS FOR THE DESTRUCTION OF NOXIOUS ANIMALS OR NOXIOUS PLANTS
- A01M7/00—Special adaptations or arrangements of liquid-spraying apparatus for purposes covered by this subclass
- A01M7/0089—Regulating or controlling systems
Landscapes
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Environmental Sciences (AREA)
- Insects & Arthropods (AREA)
- Pest Control & Pesticides (AREA)
- Wood Science & Technology (AREA)
- Zoology (AREA)
- Water Supply & Treatment (AREA)
- Soil Sciences (AREA)
- Image Processing (AREA)
Abstract
The invention provides a sorghum fertilization control method and system based on artificial intelligence, which are characterized in that an NDVI image, an NRI image and a GNDVI image are obtained according to multispectral images, and the preprocessed NDVI image is subjected to image segmentation to obtain a sorghum planting areaObtaining the sum in the NRI image and the GNDVI image according to the geographic coordinatesCorresponding sorghum planting areaWill be pretreatedAnd after pretreatmentRespectively inputting the two nitrogen fertilizer fertilizing amounts R into corresponding convolutional neural networks 1 、R 2 Respectively selectInputting the K values into the corresponding deep neural network to obtain two nitrogen fertilizer applying amounts R 3 、R 4 The method comprises the steps of carrying out a first treatment on the surface of the According toCalculating the average value of the (B) to obtain a third fertilization amount, and finally based on R 1 、R 2 、R 3 、R 4 Obtaining the third fertilizing amount and planting timeIs a fertilizer amount of nitrogen fertilizer. According to the invention, the fertilizing amount of the sorghum nitrogen fertilizer is accurately obtained by combining the vegetation indexes NDVI, RNI and GNDVI with artificial intelligence, so that the accurate control of the sorghum fertilization is realized.
Description
Technical Field
The invention relates to the field of agriculture, in particular to a sorghum fertilization control method and system based on artificial intelligence.
Background
The agricultural machine is a traditional agricultural large country, management of farmlands, such as weeding, pesticide spraying, fertilization, irrigation and the like, is completed completely by manpower, the production level is improved, and especially the application of agricultural machines such as a seeder, a harvester, an agricultural plant protection unmanned plane and the like greatly reduces the workload of a planter. The sorghum has been planted for 5000 years, belongs to the Gramineae sorghum genus, is a fifth large grain crop in the world, has the characteristics of drought tolerance, salt and alkali tolerance and the like, is widely planted in China, has three main planting areas including northeast, north China and southwest, and has the status of the sorghum continuously changed along with the improvement of living standard, and the main grain of the principle is changed into the main raw materials of brewing and feed. The advanced agricultural management technology is not available for sorghum planting, people are also continuously exploring the scientization and automation of agricultural planting, and along with the continuous maturity of image processing technology and artificial intelligence technology, more and more research institutions and enterprises begin to apply artificial intelligence to agricultural planting, and the artificial intelligence plays a role in agriculture as growth vigor and pest identification.
The nitrogenous fertilizer is a main fertilizer for sorghum, and directly affects the harvest of sorghum, and although chlorophyll meter and soil analysis can help a grower to determine the fertilizing amount, in actual production, the nitrogenous fertilizer is applied less, most growers rely on own experience to fertilize, the fertilizing amount is excessive, so that the fertilizer is wasted, the environment is polluted, the fertilizing amount is too low, and the yield of the sorghum is affected. In addition, the determination method based on visual images is also used for the fertilizing amount of the nitrogenous fertilizer, especially the application of a plant protection unmanned aerial vehicle, so that the visual images and the remote sensing images of the sorghum can be conveniently obtained, and how to accurately determine the fertilizing amount of the nitrogenous fertilizer of the sorghum by utilizing the remote sensing images shot by the unmanned aerial vehicle is a key for improving the yield of the sorghum and saving the fertilizer.
Disclosure of Invention
In order to accurately control the fertilizing amount of sorghum, on the one hand, the invention provides an artificial intelligence-based sorghum fertilizing control method, which comprises the following steps:
s1, acquiring a multispectral image of sorghum shot by a multispectral camera carried by an unmanned aerial vehicle, and acquiring an NDVI image formed by normalized vegetation indexes, an NRI image formed by nitrogen reflection indexes and a GNDVI image formed by green normalized vegetation indexes according to the multispectral image; image segmentation is carried out on the preprocessed NDVI image to obtain a sorghum planting areaWherein i is more than or equal to 1 and less than or equal to N, N is the number of the obtained sorghum planting areas;
s2, respectively obtaining the NRI image and the GNDVI image according to geographic coordinatesCorresponding sorghum planting area->Pre-treated +.>And +.>Respectively inputting the two nitrogen fertilizer fertilizing amounts R into corresponding convolutional neural networks 1 、R 2 Respectively select->Inputting the K values into the corresponding deep neural network to obtain two nitrogen fertilizer applying amounts R 3 、R 4 ;
S3, according to R 1 、R 3 Obtaining a first fertilization amount according to R 2 、R 4 Obtaining a second fertilization amount, judging the deviation of the first fertilization amount and the second fertilization amount, and if the deviation is larger than a threshold value, judging the deviation according to the following conditionsThe average value of (2) is calculated to obtain a third fertilization amount based on R 1 、R 2 、R 3 、R 4 Obtaining the area +.>Is a nitrogen fertilizer application amount; otherwise, based on R 1 、R 2 、R 3 、R 4 And planting time obtaining area +.>Is a fertilizer amount of nitrogen fertilizer.
Preferably, the image segmentation is performed on the preprocessed NDVI image to obtain a sorghum planting areaThe method comprises the following steps:
obtaining the planting time of sorghum, if the planting time is smaller than the preset time, identifying the sorghum plant and the position of the sorghum plant, deleting the non-sorghum plant part in the NDVI image, generating a second NDVI image, and dividing the second NDVI image by adopting a watershed method to obtain a sorghum planting areaOtherwise, directly adopting a watershed method to carry out image segmentation on the NDVI image to obtain a sorghum planting area +.>
Preferably, the said to be pretreatedAnd +.>Respectively input corresponding toIn the convolutional neural network, the method specifically comprises the following steps:
s22, randomly selectingMiddle W 1 ×H 1 Each pixel point is W 1 ×H 1 Mapping individual pixel points to W 1 ×H 1 Generating a first picture in the pictures; randomly select->Middle W 2 ×H 2 Each pixel point is W 2 ×H 2 Mapping individual pixel points to W 2 ×H 2 Generating a second picture in the pictures;
s23, repeating the step S22, generating a plurality of first pictures and a plurality of second pictures, putting the generated plurality of first pictures into a first picture set, and putting the generated plurality of second pictures into a second picture set; and inputting the first picture set into a first convolutional neural network, and inputting the second picture set into a second convolutional neural network.
Preferably, the selectingThe K values are input into the corresponding deep neural network, specifically:
calculation ofRespectively acquiring K/2 pixel points at the left side and the right side of the peak point, and sequentially placing the gray values of the pixel points into an array K NDVI [j]In, and array K NDVI [j]The corresponding frequency of the medium element in the gray level histogram is larger than the average value, wherein j is more than or equal to 0 and less than K;
calculation ofIs a peak point of the gray level histogram of (c),k/2 pixel points are respectively obtained at the left side and the right side of the peak point, and gray values of the pixel points are sequentially put into an array K NRI [j]In, and array K NRI [j]The corresponding frequency of the medium element in the gray level histogram is larger than the average value, wherein j is more than or equal to 0 and less than K;
setting the array K NDVI [j]Inputting a first deep neural network, and setting the array K NRI [j]A second deep neural network is input.
Preferably, if the deviation is greater than a threshold, then according toThe average value of (2) is calculated to obtain a third fertilization amount based on R 1 、R 2 、R 3 、R 4 Obtaining the area +.>The fertilizer application amount of the nitrogen fertilizer is specifically as follows:
setting an upper limit and a lower limit of the interval with the third fertilizing amount as the center, if R 1 、R 2 、R 3 、R 4 In the interval, reserving, otherwise removing, calculating the average value of reserved values
If the planting time is less than the preset time, according to the regionGeographic coordinate acquisition and region->Corresponding area in said NDVI image, calculating area +.>The ratio of the area to the area of said area, a correction factor is obtained from said ratio, for the average +.>Correcting to obtain the final productThe nitrogen fertilizer application amount in the area of the area;
Preferably, the R-based 1 、R 2 、R 3 、R 4 And planting time obtaining areaThe fertilizer application amount of the nitrogen fertilizer is specifically as follows:
if the planting time is less than the preset time, according to the regionGeographic coordinate acquisition and region->Corresponding area in said NDVI image, calculating area +.>The ratio of the area to the area of the region is used for obtaining a correction factor according to the ratio, and R is 1 、R 2 、R 3 、R 4 Mean value of>Correcting to obtain the nitrogen fertilizer application amount of the area of the region;
In addition, the invention also provides an artificial intelligence-based sorghum fertilization control system, which comprises the following modules:
image processing apparatusThe acquisition module is used for acquiring a multispectral image of sorghum shot by a multispectral camera carried by the unmanned aerial vehicle, and acquiring an NDVI image formed by normalized vegetation indexes, an NRI image formed by nitrogen reflection indexes and a GNDVI image formed by green normalized vegetation indexes according to the multispectral image; image segmentation is carried out on the preprocessed NDVI image to obtain a sorghum planting areaWherein i is more than or equal to 1 and less than or equal to N, N is the number of the obtained sorghum planting areas;
an artificial intelligence module for obtaining the NRI image and the GNDVI image according to geographic coordinatesCorresponding sorghum planting area->Pre-treated +.>And +.>Respectively inputting the two nitrogen fertilizer fertilizing amounts R into corresponding convolutional neural networks 1 、R 2 Respectively select->Inputting the K values into the corresponding deep neural network to obtain two nitrogen fertilizer applying amounts R 3 、R 4 ;
The fertilization amount control module is used for controlling the fertilization amount according to R 1 、R 3 Obtaining a first fertilization amount according to R 2 、R 4 Obtaining a second fertilization amount, judging the deviation of the first fertilization amount and the second fertilization amount, and if the deviation is larger than a threshold value, judging the deviation according to the following conditionsThe average value of (2) is calculated to obtain a third fertilization amount based on R 1 、R 2 、R 3 、R 4 Obtaining the area +.>Is a nitrogen fertilizer application amount; otherwise, based on R 1 、R 2 、R 3 、R 4 And planting time obtaining area +.>Is a fertilizer amount of nitrogen fertilizer.
Preferably, the image segmentation is performed on the preprocessed NDVI image to obtain a sorghum planting areaThe method comprises the following steps:
obtaining the planting time of sorghum, if the planting time is smaller than the preset time, identifying the sorghum plant and the position of the sorghum plant, deleting the non-sorghum plant part in the NDVI image, generating a second NDVI image, and dividing the second NDVI image by adopting a watershed method to obtain a sorghum planting areaOtherwise, directly adopting a watershed method to carry out image segmentation on the NDVI image to obtain a sorghum planting area +.>
Preferably, the said to be pretreatedAnd +.>Respectively inputting the two components into corresponding convolutional neural networks, specifically:
s22, randomly selectingMiddle W 1 ×H 1 Each pixel point is W 1 ×H 1 Mapping individual pixel points to W 1 ×H 1 Generating a first picture in the pictures; randomly select->Middle W 2 ×H 2 Each pixel point is W 2 ×H 2 Mapping individual pixel points to W 2 ×H 2 Generating a second picture in the pictures;
s23, repeating the step S22, generating a plurality of first pictures and a plurality of second pictures, putting the generated plurality of first pictures into a first picture set, and putting the generated plurality of second pictures into a second picture set; and inputting the first picture set into a first convolutional neural network, and inputting the second picture set into a second convolutional neural network.
Preferably, the selectingThe K values are input into the corresponding deep neural network, specifically:
calculation ofRespectively acquiring K/2 pixel points at the left side and the right side of the peak point, and sequentially placing the gray values of the pixel points into an array K NDVI [j]In, and array K NDVI [j]The corresponding frequency of the medium element in the gray level histogram is larger than the average value, wherein j is more than or equal to 0 and less than K;
calculation ofRespectively acquiring K/2 pixel points at the left side and the right side of the peak point, and sequentially placing the gray values of the pixel points into an array K NRI [j]In, and array K NRI [j]The corresponding frequency of the medium element in the gray level histogram is larger than the average value, wherein j is more than or equal to 0 and less than K;
setting the array K NDVI [j]Inputting a first deep neural network, and setting the array K NRI [j]A second deep neural network is input.
Preferably, if the deviation is greater than a threshold, then according toThe average value of (2) is calculated to obtain a third fertilization amount based on R 1 、R 2 、R 3 、R 4 Obtaining the area +.>The fertilizer application amount of the nitrogen fertilizer is specifically as follows:
setting an upper limit and a lower limit of the interval with the third fertilizing amount as the center, if R 1 、R 2 、R 3 、R 4 In the interval, reserving, otherwise removing, calculating the average value of reserved values
If the planting time is less than the preset time, according to the regionGeographic coordinate acquisition and region->Corresponding area in said NDVI image, calculating area +.>The ratio of the area to the area of said area, a correction factor is obtained from said ratio, for the average +.>Correcting to obtain the nitrogen fertilizer application amount of the area of the region;
Preferably, the R-based 1 、R 2 、R 3 、R 4 And planting time obtaining areaThe fertilizer application amount of the nitrogen fertilizer is specifically as follows:
if the planting time is less than the preset time, according to the regionGeographic coordinate acquisition and region->Corresponding area in said NDVI image, calculating area +.>The ratio of the area to the area of the region is used for obtaining a correction factor according to the ratio, and R is 1 、R 2 、R 3 、R 4 Mean value of>Correcting to obtain the nitrogen fertilizer application amount of the area of the region;
Finally, the invention provides a computer readable storage medium for storing computer program instructions for execution by a processor of a method as described above.
According to the invention, on the basis of NDVI and NRI vegetation indexes, a convolutional neural network and a deep neural network are adopted to calculate the nitrogen fertilizer application amount of sorghum, and whether the GNDVI vegetation indexes are adopted to correct the nitrogen fertilizer application amount is determined according to the deviation of the calculated nitrogen fertilizer application amount.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the present invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a first embodiment;
FIG. 2 is a diagram of a convolutional neural architecture in an artificial intelligence module;
FIG. 3 is a block diagram of a deep neural network in an artificial intelligence module;
fig. 4 is a schematic diagram of a gray level histogram.
Detailed Description
In this document, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
The invention provides an artificial intelligence-based sorghum fertilization control method, which is shown in fig. 1 and comprises the following steps:
s1, acquiring a multispectral image of sorghum shot by a multispectral camera carried by an unmanned aerial vehicle, and acquiring an NDVI image formed by normalized vegetation indexes, an NRI image formed by nitrogen reflection indexes and a GNDVI image formed by green normalized vegetation indexes according to the multispectral image; image segmentation is carried out on the preprocessed NDVI image to obtain a sorghum planting areaWherein i is more than or equal to 1 and less than or equal to N, N is the number of the obtained sorghum planting areas;
the multispectral camera is a camera capable of shooting more spectrums on the basis of a common camera, and then the unmanned aerial vehicle is applied to agriculture, the multispectral camera is mounted on the unmanned aerial vehicle and can shoot spectrum images of crops, and the multispectral camera is preferably a RedEdge-M, a part Sequoia or the like, or the unmanned aerial vehicle with multispectral imaging is used. The multispectral camera may capture images at multiple frequencies simultaneously, e.g., redEdge-M may capture five band images simultaneously, red, blue, green, NIR, redEdge respectively.
The vegetation index reflects the growth condition of vegetation, and common vegetation indexes include a difference vegetation index DVI (Difference Vegetation Index), a ratio vegetation index RVI (Ratio Vegetation Index), a normalized vegetation index NDVI (Normalized Difference Vegetation Index), a nitrogen reflection index NRI (Nitrogen Reflectance Index), a greenness normalized vegetation index GNDVI (Green Normalized Difference Vegetation) and the like. The correlation between NRI and NDVI and plant nitrogen is highest, the relation between NRI, NDVI and the nitrogen fertilizer fertilizing amount is explored in many researches, but the calculated nitrogen fertilizer fertilizing amount is inaccurate by using a single vegetation index, and on the basis, the invention adopts the two indexes of NDVI and NRI as main vegetation indexes of the nitrogen fertilizer fertilizing amount and GNDVI as auxiliary vegetation indexes. NCVI, NRI and GNDVIThe calculation mode of (2) is as follows: ndvi= (R NIR -R RED )/(R NIR +R RED )、NRI=(R 570 -R 670 )/(R 570 +R 670 )、GNDVI=(R 750 -R 550 )/(R 750 +R 550 )。
In a specific embodiment, the above calculation method is used to calculate the NDVI image, the NRI image, and the GNDVI image, where the pixel values are NDVI, NRI, GNDVI respectively. Due to different areas of sorghum land and different growth conditions of sorghum, the NDVI image has different values, and the NDVI image is segmented to obtain different areasNDVI values are different for different regions.
S2, respectively obtaining the NRI image and the GNDVI image according to geographic coordinatesCorresponding sorghum planting area->Pre-treated +.>And +.>Respectively inputting the two nitrogen fertilizer fertilizing amounts R into corresponding convolutional neural networks 1 、R 2 Respectively select->Inputting the K values into the corresponding deep neural network to obtain two nitrogen fertilizer applying amounts R 3 、R 4 ;
After dividing the NDVI image, a plurality of images are obtainedAt the same time, the NRI image and the GNDVI image can exist and existThe corresponding regions are respectively marked as +.>The NDVI value, NRI value and GNVDI value of the same block of sorghum land are shown.
After obtainingThen respectively inputting the nitrogen fertilizer into the corresponding pre-trained convolutional neural networks, and outputting the corresponding nitrogen fertilizer fertilization amount, wherein the specific structure is shown in figure 2; due to->The values of (2) reflect the magnitudes of NDVI and NRI, i.e. reflect the growth of sorghum, and the invention further selects +.>The K typical values of (2) are input into the corresponding deep neural network, and then the corresponding nitrogen fertilizer fertilizing amounts are output, and the specific structure is shown in figure 3.
S3, according to R 1 、R 3 Obtaining a first fertilization amount according to R 2 、R 4 Obtaining a second fertilization amount, judging the deviation of the first fertilization amount and the second fertilization amount, and if the deviation is larger than a threshold value, judging the deviation according to the following conditionsThe average value of (2) is calculated to obtain a third fertilization amount based on R 1 、R 2 、R 3 、R 4 Obtaining the area +.>Is a nitrogen fertilizer application amount; otherwise, based on R 1 、R 2 、R 3 、R 4 And planting time obtaining area +.>Is a fertilizer amount of nitrogen fertilizer.
Two nitrogenous fertilizer applying amounts are output through a convolutional neural network and a deep neural network, and the amount of nitrogenous fertilizer is increased>Two nitrogenous fertilizer applying amounts can be output through the convolutional neural network and the deep neural network, and R is as follows 1 、R 2 、R 3 、R 4 The sizes of (a) are different, and the correction is needed, wherein the correction is carried out by using GNVDI, specifically, R is judged 1 、R 2 、R 3 、R 4 If the deviation is too large, +.>And (3) calculating the average value of the obtained fertilizer to obtain a third fertilizer application amount, and correcting the third fertilizer application amount according to the third fertilizer application amount.
In different growth stages of sorghum, coverage rate is different, and when image segmentation is carried out, segmentation results are also different, for example, in a seedling stage, most of the images are bare lands, so that NDVI images are required to be processed and then segmented, specifically, the planting time of the sorghum is obtained, if the planting time is smaller than a preset time, positions of sorghum plants and sorghum plants are identified, non-sorghum plant parts in the NDVI images are deleted, a second NDVI image is generated, and a watershed method is adopted to segment the second NDVI image process to obtain a sorghum planting areaOtherwise, directly adopting a watershed method to carry out image segmentation on the NDVI image to obtain a sorghum planting area +.>
The sorghum plants and the bare land can be easily distinguished according to the NDVI value in the NDVI image, the bare land is deleted, the NDVI image only containing the plants, namely the second NDVI image, is obtained, and then the image is segmented. In a specific embodiment, filtering the second NDVI image to eliminate noise in the second NDVI image is further included before the segmentation. The preset time is determined according to the plant coverage area, for example, when sorghum grows to a certain stage, bare land is basically not seen, and the whole stage is the preset time which is in days or weeks.
After obtainingThen, the NRI image and the GNDVI image are obtained according to geographic coordinates and the +.>Corresponding sorghum planting area->Also images containing only or mostly sorghum plants.
In the image recognition or object detection, the convolutional neural network depends on the characteristics of image texture and the like, however, the image texture has no effect in the invention, and the calculation of the nitrogen fertilizer application amount is more than the pixel values from the image, namely the values of NDVI and NRI, which are subjected to pretreatmentAnd +.>Respectively inputting the two components into corresponding convolutional neural networks, specifically:
s22, randomly selectingMiddle W 1 ×H 1 Each pixel point is W 1 ×H 1 Mapping individual pixel points to W 1 ×H 1 Generating a first picture in the pictures; randomly select->Middle W 2 ×H 2 Each pixel point is W 2 ×H 2 Mapping individual pixel points to W 2 ×H 2 Generating a second picture in the pictures; in a specific embodiment, W 1 =W 2 ,H 1 =H 2 。
S23, repeating the step S22, generating a plurality of first pictures and a plurality of second pictures, putting the generated plurality of first pictures into a first picture set, and putting the generated plurality of second pictures into a second picture set; and inputting the first picture set into a first convolutional neural network, and inputting the second picture set into a second convolutional neural network.
Using random selection in generating a first picture and a second pictureAnd->The pixel points in the image can eliminate the influence of image textures on the result, and a plurality of first pictures and a plurality of second pictures are obtained through S22, thereby further utilizingAnd->The pixel value of the convolutional neural network is improved.
The deep neural network is one of neural networks, has a deeper network structure and is better in fitting. The gray histogram reflects the distribution of pixel values of the image, and in the present invention, further use is made ofCalculating the nitrogen fertilizer application amount by the histogram and the depth neural network, specifically, selecting +.>The K values are input into the corresponding deep neural network, specifically:
FIG. 4 is a schematic diagram of gray level histogram, calculatedRespectively acquiring K/2 pixel points at the left side and the right side of the peak point, and sequentially placing the gray values of the pixel points into an array K NDVI [j]In, and array K NDVI [j]The corresponding frequency of the medium element in the gray level histogram is larger than the average value, wherein j is more than or equal to 0 and less than K;
calculation ofRespectively acquiring K/2 pixel points at the left side and the right side of the peak point, and sequentially placing the gray values of the pixel points into an array K NRI [j]In, and array K NRI [j]The corresponding frequency of the medium element in the gray level histogram is larger than the average value, wherein j is more than or equal to 0 and less than K;
setting the array K NDVI [j]Inputting a first deep neural network, and setting the array K NRI [j]A second deep neural network is input.
Array K NDVI [j]Array K NRI [j]The values of (a) not only represent the concentrated areas of NDVI and NRI values, but also reflect curve change conditions, and an array K is arranged NDVI [j]Array K NRI [j]And respectively inputting the two corresponding deep neural networks to obtain two nitrogenous fertilizer applying amounts.
In addition to the above-described acquisition array K NDVI [j]Array K NRI [j]In another embodiment, a horizontal centerline of the gray histogram is obtained, a curve portion with peak points is reserved, and K numbers are obtained at equal intervals to obtain a group K NDVI [j]Array K NRI [j]。
When R is calculated 1 、R 2 、R 3 、R 4 When the difference between the two is relatively large, the obvious abnormal value needs to be removed, and if the difference is greater than the threshold value, the difference is based onAverage meter of (a)Calculating to obtain a third fertilization amount based on R 1 、R 2 、R 3 、R 4 Obtaining the area +.>The fertilizer application amount of the nitrogen fertilizer is specifically as follows: />
Setting an upper limit and a lower limit of the interval with the third fertilizing amount as the center, if R 1 、R 2 、R 3 、R 4 In the interval, reserving, otherwise removing, calculating the average value of reserved values
If the planting time is less than the preset time, according to the regionGeographic coordinate acquisition and region->Corresponding area in said NDVI image, calculating area +.>The ratio of the area to the area of said area, a correction factor is obtained from said ratio, for the average +.>Correcting to obtain the nitrogen fertilizer application amount of the area of the region;
when the planting time is less than the preset time, the area is needed to be notedRepresenting the actual planting area of the sorghum, wherein the area is formed by removing bare land; when the planting time is longer than the preset time, the area is +.>The actual planting area of the sorghum is the actual planting area.
When the planting time is shortAt a preset time, regionAreas other than the actual sorghum field need to be converted to the actual sorghum field and the results need to be corrected, because, for example, & lt, in the seedling stage area>And the jointing stage area->The same, but it is obvious that the amount of fertilizer applied in actual fertilizer application is different because the bare land in the seedling stage is more, and in view of this, when the planting time is less than the preset time, the invention further provides a mean value +.>And (5) performing correction. Specifically, calculate area->The ratio of the area to the area of said region, i.e. the calculated region +.>And the ratio of the actual sorghum planting area, based on the ratio to the average +.>Modifications are made by means including but not limited to average +.>Multiplying by said ratio, of course, +.>Multiplying the ratio is merely an example and is not limited to the above.
The third fertilization amount is calculated according toAverage value of (i.e.)>The relationship between the average value of the GNVDI region and the nitrogen fertilizer application amount is calculated, and the relationship between the GNVDI region and the nitrogen fertilizer can be obtained through clustering or data analysis, which is generally expressed in a functional mode, such as +.>The determination, of course, of the amount of nitrogen fertilizer applied by calculation based on GNVDI is only an approximation of the amount of fertilizer applied, and is not accurate, but can be used to remove R 1 、R 2 、R 3 、R 4 Is a special point of the (c).
Preferably, the R-based 1 、R 2 、R 3 、R 4 And planting time obtaining areaThe fertilizer application amount of the nitrogen fertilizer is specifically as follows:
if the planting time is less than the preset time, according to the regionGeographic coordinate acquisition and region->Corresponding area in said NDVI image, calculating area +.>The ratio of the area to the area of the region is used for obtaining a correction factor according to the ratio, and R is 1 、R 2 、R 3 、R 4 Mean value of>Correction is carried out to obtain the regionThe nitrogen fertilizer application amount in the area of the field;
In a specific embodiment, after S3, S4 is further included, according to the areaAnd (3) controlling the unmanned aerial vehicle to perform page fertilization of the nitrogen fertilizer according to the nitrogen fertilizer fertilization amount and the geographic coordinates. There are two ways to control the amount of the unmanned aerial vehicle to fertilize the page, one is to control the flying speed of the unmanned aerial vehicle, and the other is to control the spraying amount of the unmanned aerial vehicle. Specifically, the unmanned aerial vehicle flies with constant spraying amount, for the area with large nitrogen fertilizer applying amount, the unmanned aerial vehicle flies at a slower speed, and for the area with small nitrogen fertilizer applying amount, the unmanned aerial vehicle flies at a faster speed; or the unmanned aerial vehicle flies at a constant speed, and for the area with large nitrogen fertilizer fertilizing amount, the unmanned aerial vehicle flies at a larger spraying amount, and for the area with small nitrogen fertilizer fertilizing amount, the unmanned aerial vehicle flies at a smaller spraying amount.
Example two
The invention also provides an artificial intelligence-based sorghum fertilization control system, which comprises the following modules:
the image acquisition module is used for acquiring a multispectral image of sorghum shot by a multispectral camera carried by the unmanned aerial vehicle, and acquiring an NDVI image formed by normalized vegetation indexes, an NRI image formed by nitrogen reflection indexes and a GNDVI image formed by green normalized vegetation indexes according to the multispectral image; image segmentation is carried out on the preprocessed NDVI image to obtain a sorghum planting areaWherein i is more than or equal to 1 and less than or equal to N, N is the number of the obtained sorghum planting areas;
an artificial intelligence module for respectively obtaining the information according to the geographic coordinatesNRI image and GNDVI imageCorresponding sorghum planting area->Pre-treated +.>And +.>Respectively inputting the two nitrogen fertilizer fertilizing amounts R into corresponding convolutional neural networks 1 、R 2 Respectively select->Inputting the K values into the corresponding deep neural network to obtain two nitrogen fertilizer applying amounts R 3 、R 4 ;
The fertilization amount control module is used for controlling the fertilization amount according to R 1 、R 3 Obtaining a first fertilization amount according to R 2 、R 4 Obtaining a second fertilization amount, judging the deviation of the first fertilization amount and the second fertilization amount, and if the deviation is larger than a threshold value, judging the deviation according to the following conditionsThe average value of (2) is calculated to obtain a third fertilization amount based on R 1 、R 2 、R 3 、R 4 Obtaining the area +.>Is a nitrogen fertilizer application amount; otherwise, based on R 1 、R 2 、R 3 、R 4 And planting time obtaining area +.>Is a fertilizer amount of nitrogen fertilizer.
Preferably, the image segmentation is performed on the preprocessed NDVI image to obtain a sorghum planting areaThe method comprises the following steps:
obtaining the planting time of sorghum, if the planting time is smaller than the preset time, identifying the sorghum plant and the position of the sorghum plant, deleting the non-sorghum plant part in the NDVI image, generating a second NDVI image, and dividing the second NDVI image by adopting a watershed method to obtain a sorghum planting areaOtherwise, directly adopting a watershed method to carry out image segmentation on the NDVI image to obtain a sorghum planting area +.>
Preferably, the said to be pretreatedAnd +.>Respectively inputting the two components into corresponding convolutional neural networks, specifically:
s22, randomly selectingMiddle W 1 ×H 1 Each pixel point is W 1 ×H 1 Mapping individual pixel points to W 1 ×H 1 Generating a first picture in the pictures; randomly select->Middle W 2 ×H 2 Each pixel point is W 2 ×H 2 Mapping individual pixel points to W 2 ×H 2 Generating a second picture in the pictures;
s23, repeating the step S22, generating a plurality of first pictures and a plurality of second pictures, putting the generated plurality of first pictures into a first picture set, and putting the generated plurality of second pictures into a second picture set; and inputting the first picture set into a first convolutional neural network, and inputting the second picture set into a second convolutional neural network.
Preferably, the selectingThe K values are input into the corresponding deep neural network, specifically:
calculation ofRespectively acquiring K/2 pixel points at the left side and the right side of the peak point, and sequentially placing the gray values of the pixel points into an array K NDVI [j]In, and array K NDVI [j]The corresponding frequency of the medium element in the gray level histogram is larger than the average value, wherein j is more than or equal to 0 and less than K; />
Calculation ofRespectively acquiring K/2 pixel points at the left side and the right side of the peak point, and sequentially placing the gray values of the pixel points into an array K NRI [j]In, and array K NRI [j]The corresponding frequency of the medium element in the gray level histogram is larger than the average value, wherein j is more than or equal to 0 and less than K;
setting the array K NDVI [j]Inputting a first deep neural network, and setting the array K NRI [j]A second deep neural network is input.
Preferably, if the deviation is greater than a threshold, then according toThe average value of (2) is calculated to obtain a third fertilization amount based on R 1 、R 2 、R 3 、R 4 Obtaining the area +.>The fertilizer application amount of the nitrogen fertilizer is specifically as follows:
setting an upper limit and a lower limit of the interval with the third fertilizing amount as the center, if R 1 、R 2 、R 3 、R 4 In the interval, reserving, otherwise removing, calculating the average value of reserved values
If the planting time is less than the preset time, according to the regionGeographic coordinate acquisition and region->Corresponding area in said NDVI image, calculating area +.>The ratio of the area to the area of said area, a correction factor is obtained from said ratio, for the average +.>Correcting to obtain the nitrogen fertilizer application amount of the area of the region;
Preferably, the R-based 1 、R 2 、R 3 、R 4 And planting time obtaining areaThe fertilizer application amount of the nitrogen fertilizer is specifically as follows:
if the planting time is less than the preset time, according to the regionGeographic coordinate acquisition and region->Corresponding area in said NDVI image, calculating area +.>The ratio of the area to the area of the region is used for obtaining a correction factor according to the ratio, and R is 1 、R 2 、R 3 、R 4 Mean value of>Correcting to obtain the nitrogen fertilizer application amount of the area of the region;
Example III
The present invention provides a computer readable storage medium storing computer program instructions for execution by a processor of a method as in embodiment one.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by adding necessary general purpose hardware platforms, or may be implemented by a combination of hardware and software. Based on such understanding, the foregoing aspects, in essence and portions contributing to the art, may be embodied in the form of a computer program product, which 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, etc.) having computer-usable program code embodied therein.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the 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 scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (6)
1. An artificial intelligence-based sorghum fertilization control method is characterized by comprising the following steps of:
s1, acquiring a multispectral image of sorghum shot by a multispectral camera carried by an unmanned aerial vehicle, and acquiring an NDVI image formed by normalized vegetation indexes, an NRI image formed by nitrogen reflection indexes and a GNDVI image formed by green normalized vegetation indexes according to the multispectral image; image segmentation is carried out on the preprocessed NDVI image to obtain a sorghum planting areaWherein i is more than or equal to 1 and less than or equal to N, N is the number of the obtained sorghum planting areas;
s2, respectively obtaining the NRI image and the GNDVI image according to geographic coordinatesCorresponding sorghum planting area->Pre-treated +.>And +.>Respectively inputting the two nitrogen fertilizer fertilizing amounts R into corresponding convolutional neural networks 1 、R 2 Respectively select->Middle KInputting the numerical values into the corresponding deep neural network to obtain two nitrogenous fertilizer applying amounts R 3 、R 4 ;
S3, according to R 1 、R 3 Obtaining a first fertilization amount according to R 2 、R 4 Obtaining a second fertilization amount, judging the deviation of the first fertilization amount and the second fertilization amount, and if the deviation is larger than a threshold value, judging the deviation according to the following conditionsThe average value of (2) is calculated to obtain a third fertilization amount based on R 1 、R 2 、R 3 、R 4 Obtaining the area +.>Is a nitrogen fertilizer application amount; otherwise, based on R 1 、R 2 、R 3 、R 4 And planting time obtaining area +.>Is a nitrogen fertilizer application amount;
the preprocessed NDVI image is subjected to image segmentation to obtain a sorghum planting areaThe method comprises the following steps:
obtaining the planting time of sorghum, if the planting time is smaller than the preset time, identifying the sorghum plant and the position of the sorghum plant, deleting the non-sorghum plant part in the NDVI image, generating a second NDVI image, and dividing the second NDVI image by adopting a watershed method to obtain a sorghum planting areaOtherwise, directly adopting a watershed method to carry out image segmentation on the NDVI image to obtain a sorghum planting area +.>
Said pre-treatedAnd +.>Respectively inputting the two components into corresponding convolutional neural networks, specifically:
s22, randomly selectingMiddle W 1 ×H 1 Each pixel point is W 1 ×H 1 Mapping individual pixel points to W 1 ×H 1 Generating a first picture in the pictures; randomly select->Middle W 2 ×H 2 Each pixel point is W 2 ×H 2 Mapping individual pixel points to W 2 ×H 2 Generating a second picture in the pictures;
s23, repeating the step S22, generating a plurality of first pictures and a plurality of second pictures, putting the generated plurality of first pictures into a first picture set, and putting the generated plurality of second pictures into a second picture set; inputting the first picture set into a first convolutional neural network, and inputting the second picture set into a second convolutional neural network;
the respective selectionThe K values are input into the corresponding deep neural network, specifically:
calculation ofRespectively acquiring K/2 pixel points at the left side and the right side of the peak point, and sequentially placing the gray values of the pixel points into an array K NDVI [j]In, and array K NDVI [j]The corresponding frequency of the medium element in the gray level histogram is larger than the average value, wherein j is more than or equal to 0 and less than K;
calculation ofRespectively acquiring K/2 pixel points at the left side and the right side of the peak point, and sequentially placing the gray values of the pixel points into an array K NRI [j]In, and array K NRI [j]The corresponding frequency of the medium element in the gray level histogram is larger than the average value, wherein j is more than or equal to 0 and less than K; />
Setting the array K NDVI [j]Inputting a first deep neural network, and setting the array K NRI [j]A second deep neural network is input.
2. The method of claim 1, wherein if the deviation is greater than a threshold, then according toThe average value of (2) is calculated to obtain a third fertilization amount based on R 1 、R 2 、R 3 、R 4 Obtaining the area +.>The fertilizer application amount of the nitrogen fertilizer is specifically as follows:
setting an upper limit and a lower limit of the interval with the third fertilizing amount as the center, if R 1 、R 2 、R 3 、R 4 In the interval, reserving, otherwise removing, calculating the average value of reserved values
If the planting time is less than the preset time, according to the regionGeographic coordinate acquisition and region->Corresponding area in said NDVI image, calculating area +.>The ratio of the area to the area of said area, a correction factor is obtained from said ratio, for the average +.>Correcting to obtain the nitrogen fertilizer application amount of the area of the region;
3. The method of claim 1, wherein the R-based 1 、R 2 、R 3 、R 4 And planting time obtaining areaThe fertilizer application amount of the nitrogen fertilizer is specifically as follows:
if the planting time is less than the preset time, according to the regionGeographic coordinate acquisition and region->Corresponding area in said NDVI image, calculating area +.>The ratio of the area to the area of the region is used for obtaining a correction factor according to the ratio, and R is 1 、R 2 、R 3 、R 4 Mean value of>Correcting to obtain the nitrogen fertilizer application amount of the area of the region;
4. An artificial intelligence-based sorghum fertilization control system, which is characterized by comprising the following modules:
the image acquisition module is used for acquiring a multispectral image of sorghum shot by a multispectral camera carried by the unmanned aerial vehicle, and acquiring an NDVI image formed by normalized vegetation indexes, an NRI image formed by nitrogen reflection indexes and a GNDVI image formed by green normalized vegetation indexes according to the multispectral image; image segmentation is carried out on the preprocessed NDVI image to obtain a sorghum planting areaWherein i is more than or equal to 1 and less than or equal to N, N is the number of the obtained sorghum planting areas;
an artificial intelligence module for obtaining the NRI image and the GNDVI image according to geographic coordinatesCorresponding sorghum planting area->Pre-treated +.>And +.>Respectively inputting the two nitrogen fertilizer fertilizing amounts R into corresponding convolutional neural networks 1 、R 2 Respectively select->Inputting the K values into the corresponding deep neural network to obtain two nitrogen fertilizer applying amounts R 3 、R 4 ;
The fertilization amount control module is used for controlling the fertilization amount according to R 1 、R 3 Obtaining a first fertilization amount according to R 2 、R 4 Obtaining a second fertilization amount, judging the deviation of the first fertilization amount and the second fertilization amount, and if the deviation is larger than a threshold value, judging the deviation according to the following conditionsThe average value of (2) is calculated to obtain a third fertilization amount based on R 1 、R 2 、R 3 、R 4 Obtaining the area +.>Is a nitrogen fertilizer application amount; otherwise, based on R 1 、R 2 、R 3 、R 4 And planting time obtaining area +.>Is a nitrogen fertilizer application amount;
the NDVI image after pretreatment is subjected to image processingDividing to obtain sorghum planting areaThe method comprises the following steps:
obtaining the planting time of sorghum, if the planting time is smaller than the preset time, identifying the sorghum plant and the position of the sorghum plant, deleting the non-sorghum plant part in the NDVI image, generating a second NDVI image, and dividing the second NDVI image by adopting a watershed method to obtain a sorghum planting areaOtherwise, directly adopting a watershed method to carry out image segmentation on the NDVI image to obtain a sorghum planting area +.>
Said pre-treatedAnd +.>Respectively inputting the two components into corresponding convolutional neural networks, specifically:
s22, randomly selectingMiddle W 1 ×H 1 Each pixel point is W 1 ×H 1 Mapping individual pixel points to W 1 ×H 1 Generating a first picture in the pictures; randomly select->Middle W 2 ×H 2 Each pixel point is W 2 ×H 2 Mapping individual pixel points to W 2 ×H 2 Generating a second picture in the pictures;
s23, repeating the step S22, generating a plurality of first pictures and a plurality of second pictures, putting the generated plurality of first pictures into a first picture set, and putting the generated plurality of second pictures into a second picture set; inputting the first picture set into a first convolutional neural network, and inputting the second picture set into a second convolutional neural network;
the respective selectionThe K values are input into the corresponding deep neural network, specifically:
calculation ofRespectively acquiring K/2 pixel points at the left side and the right side of the peak point, and sequentially placing the gray values of the pixel points into an array K NDVI [j]In, and array K NDVI [j]The corresponding frequency of the medium element in the gray level histogram is larger than the average value, wherein j is more than or equal to 0 and less than K;
calculation ofRespectively acquiring K/2 pixel points at the left side and the right side of the peak point, and sequentially placing the gray values of the pixel points into an array K NRI [j]In, and array K NRI [j]The corresponding frequency of the medium element in the gray level histogram is larger than the average value, wherein j is more than or equal to 0 and less than K;
setting the array K NDVI [j]Inputting a first deep neural network, and setting the array K NRI [j]A second deep neural network is input.
5. The system of claim 4, wherein if the deviation is greater than a threshold, then according toThe average value of (2) is calculated to obtain a third fertilization amount based on R 1 、R 2 、R 3 、R 4 Obtaining the area +.>The fertilizer application amount of the nitrogen fertilizer is specifically as follows:
setting an upper limit and a lower limit of the interval with the third fertilizing amount as the center, if R 1 、R 2 、R 3 、R 4 In the interval, reserving, otherwise removing, calculating the average value of reserved values
If the planting time is less than the preset time, according to the regionGeographic coordinate acquisition and region->Corresponding area in said NDVI image, calculating area +.>The ratio of the area to the area of said area, a correction factor is obtained from said ratio, for the average +.>Correcting to obtain the nitrogen fertilizer application amount of the area of the region;
6. The system of claim 4, wherein the R-based 1 、R 2 、R 3 、R 4 And planting time obtaining areaThe fertilizer application amount of the nitrogen fertilizer is specifically as follows:
if the planting time is less than the preset time, according to the regionGeographic coordinate acquisition and region->Corresponding area in said NDVI image, calculating area +.>The ratio of the area to the area of the region is used for obtaining a correction factor according to the ratio, and R is 1 、R 2 、R 3 、R 4 Mean value of>Correcting to obtain the nitrogen fertilizer application amount of the area of the region; />
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211138982.1A CN115349340B (en) | 2022-09-19 | 2022-09-19 | Sorghum fertilization control method and system based on artificial intelligence |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211138982.1A CN115349340B (en) | 2022-09-19 | 2022-09-19 | Sorghum fertilization control method and system based on artificial intelligence |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115349340A CN115349340A (en) | 2022-11-18 |
CN115349340B true CN115349340B (en) | 2023-05-19 |
Family
ID=84005789
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211138982.1A Active CN115349340B (en) | 2022-09-19 | 2022-09-19 | Sorghum fertilization control method and system based on artificial intelligence |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115349340B (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2017178666A1 (en) * | 2016-04-12 | 2017-10-19 | Diego Hernan Perez Roca | Autonomous set of devices and method for detecting and identifying plant species in an agricultural crop for the selective application of agrochemicals |
JP2020054395A (en) * | 2020-01-15 | 2020-04-09 | 国立研究開発法人農業・食品産業技術総合研究機構 | Device for determining fertilizer application level and method for determining fertilizer application level |
JP2021171057A (en) * | 2020-04-20 | 2021-11-01 | 国立研究開発法人農業・食品産業技術総合研究機構 | Crop related value derivation device and crop related value derivation method |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6889620B2 (en) * | 2001-02-28 | 2005-05-10 | The Mosaic Company | Method for prescribing site-specific fertilizer application in agricultural fields |
CN102982486B (en) * | 2012-11-14 | 2015-08-12 | 北京农业信息技术研究中心 | A kind of Fertilization Decision method based on crop growing state remote sensing monitoring information |
US10664702B2 (en) * | 2016-12-30 | 2020-05-26 | International Business Machines Corporation | Method and system for crop recognition and boundary delineation |
US11263707B2 (en) * | 2017-08-08 | 2022-03-01 | Indigo Ag, Inc. | Machine learning in agricultural planting, growing, and harvesting contexts |
CN111670668B (en) * | 2020-06-05 | 2022-07-29 | 沈阳农业大学 | Accurate topdressing method for agricultural rice unmanned aerial vehicle based on hyperspectral remote sensing prescription chart |
CN112931150B (en) * | 2021-02-03 | 2022-06-03 | 华南农业大学 | Irrigation system and method based on spectral response of citrus canopy |
CN114946350A (en) * | 2022-04-18 | 2022-08-30 | 昆明学院 | Farming pertinence intelligence fertilizer injection unit |
-
2022
- 2022-09-19 CN CN202211138982.1A patent/CN115349340B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2017178666A1 (en) * | 2016-04-12 | 2017-10-19 | Diego Hernan Perez Roca | Autonomous set of devices and method for detecting and identifying plant species in an agricultural crop for the selective application of agrochemicals |
JP2020054395A (en) * | 2020-01-15 | 2020-04-09 | 国立研究開発法人農業・食品産業技術総合研究機構 | Device for determining fertilizer application level and method for determining fertilizer application level |
JP2021171057A (en) * | 2020-04-20 | 2021-11-01 | 国立研究開発法人農業・食品産業技術総合研究機構 | Crop related value derivation device and crop related value derivation method |
Non-Patent Citations (2)
Title |
---|
基于RGB颜色空间的早稻氮素营养监测研究;叶春;刘莹;李艳大;曹中盛;张丽娜;刘继忠;;中国农业大学学报(08);31-40 * |
基于ST-PCA-BP神经网络的檀香叶片全氮含量无损检测;陈珠琳;王雪峰;;应用生态学报(05);181-188 * |
Also Published As
Publication number | Publication date |
---|---|
CN115349340A (en) | 2022-11-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110909679B (en) | Remote sensing identification method and system for fallow crop rotation information of winter wheat historical planting area | |
CN110110595B (en) | Farmland image and medicine hypertrophy data analysis method based on satellite remote sensing image | |
CN113221806B (en) | Cloud platform fusion multi-source satellite image and tea tree phenological period based automatic tea garden identification method | |
US20240023536A1 (en) | Method for remediating developmentally delayed plants | |
CN109543552B (en) | Rice growth area mapping method and system based on multi-feature extraction in growth period | |
CN111418323A (en) | Nitrogen fertilizer real-time recommendation method based on facility crop canopy coverage and plant height | |
Yang et al. | Canopy structure and light interception partitioning among shoots estimated from virtual trees: comparison between apple cultivars grown on different interstocks on the Chinese Loess Plateau | |
CN113469112A (en) | Crop growth condition image identification method and system | |
CN112580491A (en) | Method and device for determining growth stage of crop and nonvolatile storage device | |
CN112001809A (en) | Method for acquiring farmland returning information of agriculture and forestry area | |
CN110598514B (en) | Method for monitoring plot scale crop seeding area of land reclamation project area | |
CN114299393A (en) | Tobacco and rice planting mode identification method based on optical and radar time sequence data | |
Moghaddam et al. | A new method in assessing sugar beet leaf nitrogen status through color image processing and artificial neural network | |
CN111178437A (en) | Fertilizing method and computer readable storage medium | |
CA3196136A1 (en) | Advanced crop manager for crops stress mitigation | |
CN115349340B (en) | Sorghum fertilization control method and system based on artificial intelligence | |
CN114821253B (en) | Method and system for regulating and controlling applicability of water and fertilizer spraying and drip irrigation integrated system | |
Zualkernan et al. | Machine Learning for Precision Agriculture Using Imagery from Unmanned Aerial Vehicles (UAVs): A Survey. Drones 2023, 7, 382 | |
CN115512233A (en) | Multi-temporal remote sensing extraction method for planting attributes of land scale cultivated land | |
CN114219227A (en) | Method and system for precise fertilization decision and plot level display | |
CN111143601A (en) | Image processing method | |
Kabir et al. | Variable fertilizer recommendation for grass production by image–based growth status | |
Chung et al. | Variable Fertilizer Recommendation by Image-based Grass Growth Status | |
CN116757867B (en) | Digital village construction method and system based on multi-source data fusion | |
CN116897670B (en) | Crop fertilization method, device, electronic equipment and storage medium |
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 |