CN103048266B - Automatic recognizing method and device for nitrogen phosphorus and potassium stress of protected tomatoes - Google Patents
Automatic recognizing method and device for nitrogen phosphorus and potassium stress of protected tomatoes Download PDFInfo
- Publication number
- CN103048266B CN103048266B CN201210528871.1A CN201210528871A CN103048266B CN 103048266 B CN103048266 B CN 103048266B CN 201210528871 A CN201210528871 A CN 201210528871A CN 103048266 B CN103048266 B CN 103048266B
- Authority
- CN
- China
- Prior art keywords
- tomato
- sample
- potassium
- nitrogen
- characteristic
- 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
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention relates to an automatic recognizing method and device for nitrogen phosphorus and potassium stress of protected tomatoes. The automatic recognizing device consists of a CCD (charge coupled device), a spectrograph, an optical system, a light source, a displacement platform, a controller, a PC (personal computer), a display and a light box. The method comprises the following steps of: firstly, performing system experiment calibration on the automatic recognizing device by a calibrating sample; then sampling a model sample during a whole growth period at equal time intervals, and collecting the reflection intensity, the polarization distribution and hyper spectrum image information of visible light-near infrared of inverted-7-shaped leaves of the tomatoes; analyzing and acquiring a main ingredient image of a nutrition stress sample according to main ingredients by combining with a measured value of chemical nitrogen phosphorus and potassium analysis and a micro structure scanning electron microscope test of the model sample, thus determining the characteristic wavelength of nitrogen phosphorus and potassium; and classifying nutrition stress characteristics into three types including a polarization characteristic, an intensity characteristic and an image characteristic, establishing the characteristic space of the nutrition stress characteristics, and establishing an automatic recognizing and comprehensive evaluation model for nitrogen phosphorus and potassium of the tomatoes according to a D-S (Dempster/Shafer) criterion theory, thus automatically recognizing and diagnosing the nitrogen phosphorus and potassium stress types of the protected tomatoes.
Description
Technical field
The present invention relates to a kind of automatic identification lossless detection method of coercing for greenhouse tomato N P and K and device, the protected crop N P and K refered in particular to based on multi-source optical information technology coerces fast automatic identification lossless detection method and device thereof.
Background technology
At present, China's facilities horticulture area has broken through 3,300,000 hectares, ranks first in the world, and serves great function to promotion rural economy.But because its yield is large, fertilizer requirement is many, soil fertility consumes high, and many areas adopt soilless culture, therefore, facility cultivation often there will be the out of proportion and nutritional deficiency symptom of the main nutrient elements such as nitrogen, phosphorus, potassium, directly affects the yield and quality of crop.Meanwhile, some producer, in order to avoid the generation of nutritional deficiency problem, excessively uses nitrogen, phosphorus, potash fertilizer, not only causes the waste of fertilizer and the pollution of area source of environment, and can cause the reduction of crop quality, the even underproduction.Therefore, in the urgent need to carrying out precise monitoring and diagnosis to main nutrient elements such as nitrogen, phosphorus, potassium in process of crop growth, the accurate management of nutrient is realized.
For a long time, the nutrient diagnosis of crop is all based on producer's experience and laboratory conventionally test.These traditional means of testing can produce crop and destroy, and affect plant growth, and the human and material resources of at substantial, poor in timeliness.Have fast based on the spectrum of blade or canopy and computer vision diagnostic techniques, facilitate, nondestructive advantage, domesticly at present in the crop alimentary Non-Destructive Testing based on EO-1 hyperion and visual pattern technology, have some correlative studys.Application number is the application for a patent for invention of CN200510088935.0, disclose lossless detection method and the surveying instrument of a kind of portable plant nitrogen and moisture, to carry out the nutrient diagnosis of plant by detecting plant leaf at the spectral reflectance strength information of four characteristic wave strong points, utilizing the inverting of four wavelength vegetation indexs to obtain nitrogen and the water percentage information of plant.Application number is a kind of method that the patent of invention of CN200710069116.0 discloses quickly non-destructive measurement for nitrogen content of tea using multiple spectrum imaging technology.Adopt the multispectral camera system of 3CCD as vision collecting device, under control of the computer, gather influences of plant crown multispectral image information by the multispectral camera system of 3CCD, can the nondestructive nitrogen nutritional status diagnosing plant.Application number is the application for a patent for invention of CN201010208851.7, disclose a kind of tea leaf nitrogen based on high light spectrum image-forming technology, phosphorus, the quick detecting method of potassium and device, based on the high light spectrum image-forming technology of filter type, obtain the n p k nutrition information of tealeaves blade, the nutrition Non-Destructive Testing of growth of tea plant process can be realized.At present, computer vision technique mainly utilizes the Macroscopic physical characteristics such as the color of blade or canopy, texture, form to diagnose, spectroscopic diagnostics technology mainly utilizes the change of the spectral reflectivity of characteristic wave strong point to carry out the nutrition condition of inverting crop, fully cannot characterize characteristic information abundant when crop alimentary wanes, thus constrain the application of crop alimentary method for quick.The information of light wave is enriched very much, comprise intensity, wavelength, position phase and polarization state, and at present both at home and abroad Nutrition monitoring with only light wave intensity (i.e. reflectivity or reflection strength) and wavelength information, i.e. the disease conditions of employing reflectance spectrum technology, image technique or multispectral image diagnosis of technique crop.This also affects and constrains the raising of diagnostic accuracy of crop pest.Polarization image has the advantage not available for normal image and reflectance spectrum, the information that some intensity images and spectrum are difficult to characterize can be characterized, as the change of the microstructure change of target surface, the inner selective absorbing to incident light of material, scattering and body surface forward reflection, retroreflection, diffusing characteristic diffuser.Due to the unique distinction that polarization imaging technology has, therefore can coerce to crop alimentary the blade surface quality that causes and microstructure change information is extracted and characterized.The present invention is by building multi-source optical information harvester, the Typical Representative tomato of facility fruit and vegetable is detected object, obtain reflection strength, polarization, the visual spectrum Multi-Source Integration information of greenhouse tomato blade, the feature such as color, texture, metamorphosis that can cause tomato nutritional deficiency carries out visual analyzing, can detect again quality and the polarization state of microstructure, the anisotropic elastic solid information of reflection strength that nutritional deficiency causes.Effectively can improve accuracy of identification and stability that greenhouse tomato n p k nutrition coerces.At present, there is not yet Patents and report both at home and abroad.
Summary of the invention
The object of this invention is to provide a kind of greenhouse tomato N P and K and coerce automatic identifying method and device.By the multi-source optical information harvester built voluntarily, obtain the Multi-Source Integration information of greenhouse tomato, extract intensity, polarization, HYPERSPECTRAL IMAGERY feature that n p k nutrition is coerced, based on multisource information fusion technology, set up automatic recognition classification and the evaluation model of greenhouse tomato Nutrient Stress, realize waning the automatic identification of kind and quick diagnosis to greenhouse tomato n p k nutrition.For under facility condition, accurately the using of nutrient solution of tomato production process provides scientific basis.
A kind of greenhouse tomato N P and K of the present invention is coerced automatic identification equipment and is comprised as lower component: CCD, spectrograph, optical system, light source, displacement platform, controller, PC, display and light box.The effect of light box is shielding external interference, stable light source and testing environment is provided for multi-source optical information detects, light box top center secures the multi-source Sensor of Optical Information be made up of CCD, spectrograph and optical system, the bottom surface geometric center of light box secures displacement platform, multi-source Sensor of Optical Information is positioned at directly over displacement platform, perpendicular to displacement platform, in the middle part of light box, left and right sides symmetry has installed light source; Wherein CCD, spectrograph are connected with controller by data line with light source, PC is connected by data line with controller, controller accepts the steering order of PC, the parameters such as the sweep velocity when control intensity of light source, the gait of march of displacement platform and multi-source information acquiring, time shutter, focal length and rotatory polarization sheet position, implementation information collection controls; Display is for implementing collection and the control information of monitoring PC.
Wherein said CCD, spectrograph and optical system together constitute multi-source Sensor of Optical Information, and it is optical system bottom, upwards connects spectrograph, CDD successively; Wherein said optical system comprises rotatory polarization sheet, camera lens and filter, and rotatory polarization sheet can adjust polarization angle manually or automatically; Wherein said CCD comprises Visible-light CCD and Near Infrared CCD, and Visible-light CCD areas imaging is 400-1100nm, and Near Infrared CCD is indium gallium arsenic imaging CCD, and spectral range is 900-1700nm.
Wherein said light source is halogen tungsten lamp light source, and wavelength coverage is 400-2600nm.
Wherein said light box inside adopts black ESD spraying.
For realizing the object of invention, a kind of greenhouse tomato N P and K of the present invention is coerced automatic identifying method and is carried out according to following step:
1) system calibrating: tomato is demarcated sample and be fixed on a kind of greenhouse tomato N P and K and coerce on the displacement platform of automatic identification equipment, carry out sampling rating test, determine the best velocity of displacement and the polarizing angle of optical system, the imaging focal length of CCD and the time shutter that make the distortionless displacement platform of scan image; Gather Hei Chang and white field information, obtain the relative reference value of different wave length point, using the difference of the Bai Chang of each wavelength points and black field as denominator, calculate relative reflectance and the imaging gray-scale value of each pixel.
2) information acquisition: in the whole growth period of tomato, the collection of a multi-source optical information is carried out at interval of 10 days, choose each nitrogen, phosphorus, Potassium Levels and mutual test sample 20 strain at every turn, obtain respectively different nitrogen, phosphorus, Potassium Levels tomato leaf as model sample; Gather tomato leaf in the visible ray of 400-1700nm and the distribution of near infrared reflection strength, degree of polarization distribution, HYPERSPECTRAL IMAGERY information; Flow Analyzer is adopted to measure the total nitrogen of blade, total phosphorus, total potassium; Non-smooth surface characteristic is measured by scanning electron microscope and micro-image method, and the change of the internal organizational structure such as pore, cavernous body, palisade tissue.
3) feature extraction and compensation: based on the multi-source information obtained, in conjunction with the chemical measured value of N P and K, utilizes principal component analysis (PCA) to obtain the major component image of Nutrient Stress sample, determines the characteristic wavelength distribution of N P and K; On this basis, application Virtual Lab software obtains the degree of polarization feature such as degree of polarization intensity distributions, Stocks parameter, Mueller matrix of characteristic wave strong point, the reflection strength distribution characteristics of the visual spectrum in application ENVI software decimation target area; The characteristic image extracting each characteristic wave strong point Nutrient Stress sample extracts its gray scale, texture, area features; Adopt orthogonal test, testing and analysis different nitrogen, phosphorus, Potassium Levels and moisture level are on polarization image, spectral reflectance intensity distributions, the intensity profile of ultraphotic spectrum signature image and the impact of textural characteristics, with the reflection strength of 489nm wavelength, blade total reducing sugar is described, analyze the Changing Pattern of different phosphate trophic level to nitrogen sugar ratio, provide the modifying factor of nitrogen diagnosis; Describe the moisture of blade with the reflection strength of 1450nm wavelength, calculate the modifying factor of nutrient and moisture respectively; Utilize each modifying factor to coerce feature correction to n p k nutrition, to reduce between N P and K and and moisture between interactive impact.
4) foundation of model of cognition: the Nutrient Stress characteristic information of tomato is divided into polarization characteristic, strength characteristic and characteristics of image three class, set up respective feature space respectively, three neural network classifiers are adopted to carry out identification that N P and K coerces and classification at respective feature space respectively, three neural networks are by after training and learning to obtain respective recognition result, utilize D-S criterion theoretical, combined standard control group sample, sets up automatic identification and the comprehensive evaluation model of the tomato N P and K of Different growth phases; Based on the built-up pattern obtained, in any self-sow stage of greenhouse tomato, what gather tomato falls 7 leaves as detection sample, multi-source optical information harvester is utilized to gather the multi-source feature of tomato, and input tomato nutrient and coerce automatic recognition software, select corresponding growth phase and output intent option, operating software identifies nitrogen, phosphorus, the stress state of potassium, the trophic level of tomato automatically, and provides corresponding nutrient solution allotment control strategy.
Wherein said tomato sample is divided into be demarcated sample, model sample and detects sample.The sample of employing when demarcation sample refers to and carries out system calibrating, takes from 7 leaves that fall of standard control group plant, the sample that model sample uses when referring to and set up the automatic model of cognition of N P and K, for ensureing the accuracy of the pure and model of cognition of sample, this sample adopts greenhouse nutrient solutions cultivation mode, select the mode that solid matrix groove is trained, and divide nitrogen, phosphorus, tomato model sample is cultivated in potassium test site, the different nitrogen of drip irrigation at regular time and quantity in each test site, phosphorus, the nutrient solution of potassium concentration, nutrient solution prescription adopts the rugged formula in mountain, in normal recipe, according to 25% of this element normal contents, 50%, 75%, 100%, 150%, 200% irrigates, form different nitrogen, phosphorus, the single nutritional deficiency sample of Potassium Levels, adopt normal orthogonal experimental design, cultivate nitrogen, phosphorus, potassium interaction model sample, detect the measurement sample that sample refers to random selecting, selected leaf position and model sample are down 7 leaves, but detect the greenhouse tomato sample that sample is nature plantation, do not carry out special classification process to nutrient solution prescription.
The wherein said automatic identification of setting up the tomato N P and K of Different growth phases and comprehensive evaluation model, refer to the time interval order gathered according to multi-source optical information, set up the model of cognition of each timing node respectively, form built-up pattern storehouse, and then utilizing the multi-source information in all stages, the n p k nutrition in matching whole growth period coerces automatic model of cognition.Wherein said combined standard control group sample, refer to that N P and K when being cultivated by sample normally uses sample, namely as a control group, the N-P-K content obtained with its chemical analysis contrasts for standard recipe and the tomato sample under managing, and obtains the trophic level detecting sample.
beneficial effect of the present invention:(1) the present invention adopts intensity, polarization, HYPERSPECTRAL IMAGERY multi-source optical information carry out the automatic recognition detection that greenhouse tomato N P and K is coerced, and the kind wane to N P and K and degree differentiate, this does not all relate in file in the past.(2) the present invention is by the synchronous Multi-Source Integration information obtaining tomato leaf, the features such as Fusion of Color, texture, form, microstructure carry out identification and the diagnosis of tomato nutritional deficiency kind, achieve automatic identification and the evaluation of tomato N P and K, the precision and stability comparing identification with traditional reflectance spectrum and the single detection meanss such as visual pattern is significantly improved.
Accompanying drawing explanation
Fig. 1 is that a kind of greenhouse tomato N P and K of the present invention coerces automatic identification equipment structural representation;
1-CCD; 2-spectrograph; 3-optical system; 4-light source; 5-displacement platform; 6-controller; 7-PC machine;
8-display; 9-light box.
Fig. 2 is that a kind of greenhouse tomato N P and K of the present invention coerces automatic identifying method process flow diagram.
Embodiment
Below in conjunction with accompanying drawing, 1 couple of the present invention is explained in further detail.
A kind of greenhouse tomato N P and K of the present invention is coerced automatic identification equipment and is comprised as lower component: CCD1, spectrograph 2, optical system 3, light source 4, displacement platform 5, controller 6, PC 7, display 8 and light box 9.The effect of light box 9 is shielding external interference, stable light source and testing environment is provided for multi-source optical information detects, light box 9 top center secures the multi-source Sensor of Optical Information be made up of CCD1, spectrograph 2 and optical system 3, the bottom surface geometric center of light box secures displacement platform 5, multi-source Sensor of Optical Information is positioned at directly over displacement platform 5, perpendicular to displacement platform 5, in the middle part of light box 9, left and right sides symmetry has installed light source 4; Wherein CCD1, spectrograph 2 are connected with controller 6 by data line with light source 3, PC 7 is connected by data line with controller 6, controller 6 accepts the steering order of PC 7, the parameters such as the sweep velocity when control intensity of light source 4, the gait of march of displacement platform 5 and multi-source information acquiring, time shutter, focal length and rotatory polarization sheet position, implementation information collection controls; Display 8 is for implementing collection and the control information of monitoring PC 7.
Wherein said CCD1, spectrograph 2 and optical system 3 together constitute multi-source Sensor of Optical Information, and it is optical system 3 bottom, upwards connect spectrograph 2, CDD1 successively thereafter; Wherein said optical system 3 comprises rotatory polarization sheet, camera lens and filter, and rotatory polarization sheet can adjust polarization angle manually or automatically; Wherein said CCD1 comprises Visible-light CCD and Near Infrared CCD, and Visible-light CCD areas imaging is 400-1100nm, and Near Infrared CCD is indium gallium arsenic imaging CCD, and spectral range is 900-1700nm.
Wherein said light source be halogen tungsten lamp light source, wavelength coverage is 400-2600nm.
Wherein said light box inside adopts black ESD spraying.
Be detected as example with seedling tomato N P and K below and introduce concrete detection method.During actual measurement, carry out according to following step:
1) system calibrating: what gather tomato falls 7 leaves as seedling tomato sample, chooses 85 samples altogether, wherein sample is demarcated in 5 conducts, and other 80 as detecting sample.First demarcation sample being fixed on a kind of greenhouse tomato N P and K coerces on the displacement platform 5 of automatic identification equipment, first sampling rating test is carried out, through actual scanning, determine that the best displacement speed setting value making the distortionless displacement platform of scan image 5 is 9, the polarizing angle of optical system is 45 °, when to select the imaging focal length of CCD to be the object distance of 12mm(distance of camera lens blade be 70cm), shutter speed is 1/1000s; Gather Hei Chang and white field information, wherein black field is undertaken by closedown light source and lens cap scanning, white field is undertaken by scanning standard blank, obtain the relative reference value of 400-1700nm wavelength coverage, wherein the reference value of black field is 0, the relative reference value of white field be the difference of 4000, Yi Baichang and black field as denominator, calculate relative reflectance and the imaging gray-scale value of each pixel.
2) information acquisition: choose each not nitrogen, phosphorus, Potassium Levels and mutual test sample 80 strain, obtain respectively different nitrogen, phosphorus, Potassium Levels tomato leaf as model sample; Gather tomato leaf in the visible ray of 400-1700nm and the distribution of near infrared reflection strength, degree of polarization distribution, HYPERSPECTRAL IMAGERY; The total nitrogen of Flow Analyzer to blade is adopted to measure; The internal organizational structure such as sample surface characteristic and pore is measured by scanning electron microscope and micro-image method.
3) feature extraction and compensation: based on the multi-source information obtained, in conjunction with the chemical measured value of N P and K, utilizes principal component analysis (PCA) to obtain the major component image of Nutrient Stress sample, determines the characteristic wavelength distribution of N P and K; On this basis, application Virtual Lab software obtains the degree of polarization feature such as degree of polarization intensity distributions, Stocks parameter, Mueller matrix of characteristic wave strong point, the reflection strength distribution characteristics of the visual spectrum in application ENVI software decimation target area; The characteristic image extracting each characteristic wave strong point Nutrient Stress sample extracts its gray scale, texture, area features; Adopt orthogonal test, testing and analysis different nitrogen, phosphorus, Potassium Levels and moisture level are on polarization image, spectral reflectance intensity distributions, the intensity profile of ultraphotic spectrum signature image and the impact of textural characteristics, with the reflection strength of 489nm wavelength, blade total reducing sugar is described, analyze the Changing Pattern of different phosphate trophic level to nitrogen sugar ratio, provide the modifying factor of nitrogen diagnosis; Describe the moisture of blade with the reflection strength of 1450nm wavelength, calculate the modifying factor of nutrient and moisture respectively; Utilize each modifying factor to coerce feature correction to n p k nutrition, to reduce between N P and K and and moisture between interactive impact.
4) foundation of model of cognition: the Nutrient Stress characteristic information of tomato is divided into polarization characteristic, strength characteristic and characteristics of image three class, set up respective feature space respectively, three neural network classifiers are adopted to carry out identification that N P and K coerces and classification at respective feature space respectively, three neural networks are by after training and learning to obtain respective recognition result, utilize D-S criterion theoretical, combined standard control group sample, sets up automatic identification and the comprehensive evaluation model of the tomato N P and K of Different growth phases.Wherein, three neural network classifiers select the input number of nodes of networks according to the number of every category feature, choose the input as three sub neural networks of 7 polarization characteristic variablees, 5 strength characteristic variablees and 8 characteristics of image variablees respectively; Adopt 3 layer network structures of single hidden layer, hidden layer unit number is 10; The learning rate of training is 0.47, and learning error is 0.01, and maximum frequency of training is 2000 times.Each neural network carries out the recognition and classification of sample respectively in respective feature space, after obtaining respective recognition result, utilizes D-S criterion theory to carry out fusion decision-making and judges.
Based on the built-up pattern obtained, in any self-sow stage of greenhouse tomato, what gather tomato falls 7 leaves as detection sample, multi-source optical information harvester is utilized to gather the multi-source feature of tomato, and input tomato nutrient and coerce automatic recognition software, select corresponding growth phase and output intent option, operating software identifies nitrogen, phosphorus, the stress state of potassium, the trophic level of tomato automatically.The correct recognition rata that wherein Nitrogen In Tomato wanes reaches 100%, and the correct recognition rata of different nitrogen level reaches more than 92%; The correct recognition rata that tomato phosphorus wanes reaches 86%, and the correct recognition rata of different P levels reaches more than 80%, and the correct recognition rata that tomato potassium wanes is 70%.And research is in the past to the qualitative analysis only can done with phosphorus and potassium element, diagnosis and the accuracy of identification of nitrogen are also significantly increased.Result shows, utilizes many senses information mix together technology to carry out the quantitative test of rape nitrogen and water percentage, and as compared to single detection methods such as spectrum, visual pattern and canopy surface temperatures, precision of prediction is significantly improved.
Claims (6)
1. greenhouse tomato N P and K coerces an automatic identifying method, it is characterized in that, the step comprised is:
1) system calibrating: tomato is demarcated sample and be fixed on a kind of greenhouse tomato N P and K and coerce on the displacement platform of automatic identification equipment, carry out sampling rating test, determine the best velocity of displacement and the polarizing angle of optical system, the imaging focal length of CCD and the time shutter that make the distortionless displacement platform of scan image; Gather Hei Chang and white field information, obtain the relative reference value of different wave length point, using the difference of the Bai Chang of each wavelength points and black field as denominator, calculate relative reflectance and the imaging gray-scale value of each pixel;
2) information acquisition: in the whole growth period of tomato, the collection of a multi-source optical information is carried out at interval of 10 days, choose each nitrogen, phosphorus, Potassium Levels and mutual test sample 20 strain at every turn, obtain respectively different nitrogen, phosphorus, Potassium Levels tomato leaf as model sample; Described light source is halogen tungsten lamp light source, and wavelength coverage is 400-2600nm; What gather tomato falls 7 leaves in the visible ray of 400-1700nm and the distribution of near infrared reflection strength, degree of polarization distribution, HYPERSPECTRAL IMAGERY information; Flow Analyzer is adopted to measure the total nitrogen of blade, total phosphorus, total potassium; Non-smooth surface characteristic is measured by scanning electron microscope and micro-image method, and the change of the internal organizational structure such as pore, cavernous body, palisade tissue;
3) feature extraction and compensation: based on the multi-source information obtained, in conjunction with the chemical measured value of N P and K, utilizes principal component analysis (PCA) to obtain the major component image of Nutrient Stress sample, determines the characteristic wavelength distribution of N P and K; On this basis, the degree of polarization feature of the degree of polarization intensity distributions of Virtual Lab software acquisition characteristic wave strong point, Stocks parameter, Mueller matrix is applied, the reflection strength distribution characteristics of the visual spectrum in application ENVI software decimation target area; The characteristic image extracting each characteristic wave strong point Nutrient Stress sample extracts its gray scale, texture, area features; Adopt orthogonal test, testing and analysis different nitrogen, phosphorus, Potassium Levels and moisture level are on polarization image, spectral reflectance intensity distributions, the intensity profile of ultraphotic spectrum signature image and the impact of textural characteristics, with the reflection strength of 489nm wavelength, blade total reducing sugar is described, analyze the Changing Pattern of different phosphate trophic level to nitrogen sugar ratio, provide the modifying factor of nitrogen diagnosis; Describe the moisture of blade with the reflection strength of 1450nm wavelength, calculate the modifying factor of nutrient and moisture respectively; Utilize each modifying factor to coerce feature correction to n p k nutrition, to reduce between N P and K and and moisture between interactive impact;
4) foundation of model of cognition: the Nutrient Stress characteristic information of tomato is divided into polarization characteristic, strength characteristic and characteristics of image three class, set up respective feature space respectively, three neural network classifiers are adopted to carry out identification that N P and K coerces and classification at respective feature space respectively, three neural networks are by after training and learning to obtain respective recognition result, utilize D-S criterion theoretical, combined standard control group sample, sets up automatic identification and the comprehensive evaluation model of the tomato N P and K of Different growth phases; Based on the model obtained, in any self-sow stage of greenhouse tomato, what gather tomato falls 7 leaves as detection sample, multi-source optical information harvester is utilized to gather the multi-source feature of tomato, and input tomato nutrient and coerce automatic recognition software, select corresponding growth phase and output intent option, operating software identifies nitrogen, phosphorus, the stress state of potassium, the trophic level of tomato automatically, and provides corresponding nutrient solution allotment control strategy.
2. a kind of greenhouse tomato N P and K according to claim 1 coerces automatic identifying method, it is characterized in that, step 2) described model sample employing greenhouse nutrient solutions cultivation mode, select the mode that solid matrix groove is trained, and divide nitrogen, phosphorus, tomato model sample is cultivated in potassium test site, the different nitrogen of drip irrigation at regular time and quantity in each test site, phosphorus, the nutrient solution of potassium concentration, nutrient solution prescription adopts the rugged formula in mountain, in normal recipe, according to 25% of this element normal contents, 50%, 75%, 100%, 150%, 200% irrigates, form different nitrogen, phosphorus, the single nutritional deficiency sample of Potassium Levels, adopt normal orthogonal experimental design, cultivate nitrogen, phosphorus, potassium interaction model sample.
3. a kind of greenhouse tomato N P and K according to claim 1 and 2 coerces automatic identifying method, it is characterized in that, step 4) described in the automatic identification of setting up the tomato N P and K of Different growth phases and comprehensive evaluation model, refer to the time interval order gathered according to multi-source optical information, set up the model of cognition of each timing node respectively, form built-up pattern storehouse, and then utilize the multi-source information in all stages, the n p k nutrition in matching whole growth period coerces automatic model of cognition; Described combined standard control group sample, refer to that N P and K when being cultivated by sample normally uses sample, namely as a control group, the N-P-K content obtained with its chemical analysis contrasts for standard recipe and the tomato sample under managing, and obtains the trophic level detecting sample.
4. a kind of greenhouse tomato N P and K implemented the claims described in 1 coerces the device of automatic identifying method, comprise as lower component: CCD, spectrograph, optical system, light source, displacement platform, controller, PC, display and light box, the multi-source Sensor of Optical Information be made up of described CCD, spectrograph and optical system is fixed in light box inner top center, in light box bottom surface geometric center fix described displacement platform; Described multi-source Sensor of Optical Information is positioned at directly over displacement platform, perpendicular to displacement platform; Source symmetric is arranged on the left and right sides at middle part in light box; Wherein said CCD, spectrograph are connected with controller by data line with light source, PC is connected by data line with controller, controller accepts the steering order of PC, the parameters such as sweep velocity when controlling the intensity of described light source, the gait of march of displacement platform and multi-source information acquiring, time shutter, focal length and rotatory polarization sheet position, implementation information collection controls; Display is for implementing collection and the control information of monitoring PC.
5. a kind of greenhouse tomato N P and K coerces automatic identification equipment according to claim 4, it is characterized in that, described optical system is positioned at multi-source Sensor of Optical Information bottom, upwards connects spectrograph, CDD successively; Described optical system comprises rotatory polarization sheet, camera lens and filter, and rotatory polarization sheet can adjust polarization angle manually or automatically; Described CCD comprises Visible-light CCD and Near Infrared CCD, and Visible-light CCD areas imaging is 400-1100nm, and Near Infrared CCD is indium gallium arsenic imaging CCD, and spectral range is 900-1700nm.
6. according to claim 4 or 5, a kind of greenhouse tomato N P and K coerces automatic identification equipment, it is characterized in that, described light box inside adopts black ESD spraying.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210528871.1A CN103048266B (en) | 2012-12-11 | 2012-12-11 | Automatic recognizing method and device for nitrogen phosphorus and potassium stress of protected tomatoes |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210528871.1A CN103048266B (en) | 2012-12-11 | 2012-12-11 | Automatic recognizing method and device for nitrogen phosphorus and potassium stress of protected tomatoes |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103048266A CN103048266A (en) | 2013-04-17 |
| CN103048266B true CN103048266B (en) | 2015-06-10 |
Family
ID=48060985
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210528871.1A Active CN103048266B (en) | 2012-12-11 | 2012-12-11 | Automatic recognizing method and device for nitrogen phosphorus and potassium stress of protected tomatoes |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103048266B (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103335948B (en) * | 2013-05-31 | 2016-05-11 | 合肥众沃仪器技术有限公司 | Intensity of press marks on cigarette paper measurement mechanism and method |
| CN104198397B (en) * | 2014-07-30 | 2017-09-05 | 江苏大学 | The method that chamber crop nutrient content is detected under N P and K reciprocation |
| CN104502285A (en) * | 2014-12-21 | 2015-04-08 | 华东交通大学 | Device and method for detecting content of protein in tomato leaves by multi-angle in situ spectrum |
| CN104865194A (en) * | 2015-04-03 | 2015-08-26 | 江苏大学 | Detection apparatus and method for pesticide residues in vegetable based on near infrared, fluorescence and polarization multi-spectrum |
| US10126234B2 (en) * | 2015-10-23 | 2018-11-13 | Panasonic Intellectual Property Management Co., Ltd. | Water content of a part of plant evaluation method and water content of a part of plant evaluation apparatus |
| CN105548113B (en) * | 2015-12-31 | 2019-04-16 | 浙江大学 | A kind of plant physiology monitoring method based on chlorophyll fluorescence and multispectral image |
| CN107290293A (en) * | 2017-06-06 | 2017-10-24 | 浙江大学 | A kind of spectral imaging technology monitors the device of water planting heavy metal stress crop on-line |
| CN108323295B (en) * | 2017-12-05 | 2019-12-03 | 江苏大学 | A kind of seedling stage crop liquid manure based on multiple dimensioned habitat information detects and controls method and device |
| CN108387262B (en) * | 2018-01-03 | 2020-05-01 | 江苏大学 | Greenhouse information automatic monitoring method based on suspension type sliding rail platform |
| CN109632688B (en) * | 2018-11-28 | 2021-05-28 | 北京农业智能装备技术研究中心 | Plant seedling nutrient profit and loss state identification method and spectrum detection system |
| CN109459391B (en) * | 2019-01-07 | 2021-07-23 | 塔里木大学 | Red date quality detection and red date polarization detection model generation method and device |
| CN110119086B (en) * | 2019-04-19 | 2022-03-18 | 淮阴工学院 | Tomato greenhouse environmental parameter intelligent monitoring device based on ANFIS neural network |
| CN111220734B (en) * | 2020-02-26 | 2021-06-22 | 江苏大学 | Method for evaluating nutritional state of crops in latent nutrient deficiency period by using sensitive metabolites |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101936882A (en) * | 2010-08-10 | 2011-01-05 | 江苏大学 | Nondestructive testing method and device for nitrogen and water of crops |
| WO2011102739A1 (en) * | 2010-01-21 | 2011-08-25 | Mackenzie Research Group Limited | Improvements in and relating to methods and apparatus for applying substances to an area of interest |
| CN102384767A (en) * | 2011-11-17 | 2012-03-21 | 江苏大学 | Nondestructive detection device and method for facility crop growth information |
| CN102789579A (en) * | 2012-07-26 | 2012-11-21 | 同济大学 | Identification method for stressed state of water fertilizer of greenhouse crop on basis of computer vision technology |
-
2012
- 2012-12-11 CN CN201210528871.1A patent/CN103048266B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011102739A1 (en) * | 2010-01-21 | 2011-08-25 | Mackenzie Research Group Limited | Improvements in and relating to methods and apparatus for applying substances to an area of interest |
| CN101936882A (en) * | 2010-08-10 | 2011-01-05 | 江苏大学 | Nondestructive testing method and device for nitrogen and water of crops |
| CN102384767A (en) * | 2011-11-17 | 2012-03-21 | 江苏大学 | Nondestructive detection device and method for facility crop growth information |
| CN102789579A (en) * | 2012-07-26 | 2012-11-21 | 同济大学 | Identification method for stressed state of water fertilizer of greenhouse crop on basis of computer vision technology |
Non-Patent Citations (1)
| Title |
|---|
| 《不同生育期氮磷钾胁迫对菊花黄酮类化合物的代谢调控研究》;刘伟;《华中农业大学博士学位论文》;20101229;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103048266A (en) | 2013-04-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103048266B (en) | Automatic recognizing method and device for nitrogen phosphorus and potassium stress of protected tomatoes | |
| CN101936882B (en) | Nondestructive testing method and device for nitrogen and water of crops | |
| Hall et al. | Optical remote sensing applications in viticulture‐a review | |
| CN101915738B (en) | Method and device for rapidly detecting nutritional information of tea tree based on hyperspectral imaging technique | |
| CN104198396B (en) | The method that polarization-hyperspectral technique diagnosis crop N P and K wanes | |
| CN102788752B (en) | Non-destructive detection device and method of internal information of crops based on spectrum technology | |
| CN107796764B (en) | Method for constructing wheat leaf area index estimation model based on three-band vegetation index | |
| CN202382768U (en) | Facility crop growth information nondestructive test apparatus | |
| CN102564962B (en) | System and method for detecting nutritional components in fruit trees | |
| CN102081039A (en) | Environment-controllable hyperspectral image detecting device for crop nutrition and moisture | |
| CN104614321B (en) | A kind of crop growing state method of real-time based on spectrum picture | |
| CN103018196A (en) | Fast detection method for rape water demand information | |
| CN102445421B (en) | Preparation method of leaf colour card used for rapidly diagnosing nitrogen of flue-cured tobacco | |
| CN103439265A (en) | Real-time monitoring method for growth characters of tea trees in intensive cultivation | |
| CN106841051A (en) | A kind of crop nitrogen nutrition detection method based on visual image fusion value | |
| CN106525731A (en) | Canopy-leaf-nitrogen vertical distribution detection method and device based on remote sensing and agronomy knowledge | |
| CN105136732A (en) | Field crop dual band imaging NDVI measurement apparatus | |
| CN109187417A (en) | A kind of Terahertz of UAV system involves high-spectrum remote-sensing crops monitoring system | |
| CN103278503A (en) | Multi-sensor technology-based grape water stress diagnosis method and system therefor | |
| CN104236486A (en) | Rapid lossless measuring method for cotton leaf area index | |
| CN102788796B (en) | Nutrient diagnosis device and nutrient diagnosis method for nitrogen of crops based on multi-information integration of high spectral images and fluorescent images | |
| CN112816441B (en) | Method and device for detecting growth condition of facility horticultural crop | |
| Guillén-Climent et al. | Estimating radiation interception in an olive orchard using physical models and multispectral airborne imagery | |
| CN106404699A (en) | Nondestructive measurement method for contents of nitrogen elements of pear tree leaf blades | |
| CN202024959U (en) | Uniformization vegetation indifference index-based wireless nitrogen nutrition detecting instrument |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |