CN103940748B - Based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique - Google Patents
Based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique Download PDFInfo
- Publication number
- CN103940748B CN103940748B CN201410084466.4A CN201410084466A CN103940748B CN 103940748 B CN103940748 B CN 103940748B CN 201410084466 A CN201410084466 A CN 201410084466A CN 103940748 B CN103940748 B CN 103940748B
- Authority
- CN
- China
- Prior art keywords
- nitrogen content
- oranges
- tangerines
- canopy
- hyperspectral
- 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 discloses the prediction of a kind of oranges and tangerines canopy nitrogen content based on hyperspectral technique and visualization method, find optimal bands combined by the correlation analysis of test figure and set up inverse model, oranges and tangerines canopy leaves nitrogen content low cost can be realized, detect fast; Empirically selected characteristic wave band has more science than ever mutually, and the distribution of oranges and tangerines canopy leaves nitrogen content is carried out visual expression directly perceived by the present invention, significant for oranges and tangerines orchard site-specific nutrient management; Can according to the actual needs, utilize the method to select other vegetation index or other nutrient or other plant, set up corresponding model.
Description
Technical field
The invention belongs to agricultural technology field, particularly relate to the prediction of a kind of oranges and tangerines canopy nitrogen content based on hyperspectral technique and visualization method.
Background technology
Nitrogen is the necessary element in growth and development of fruit tree process, and it not only participates in the synthesis of many important compound such as protein, chlorophyll and enzyme directly, and can affect indirectly the metabolism of fruit tree by affecting photosynthesis.Therefore, detect fruit tree nitrogen level timely and accurately and can provide effective information for the formulation of fruit tree quantitative fertilization scheme.These precision agriculture control measures not only can ensure the output efficiency in fruit quality and orchard, and can relax the water resource pollution problem caused by excessive nitrogenous fertilizer.
Oranges and tangerines are one of main fruit trees of extensively plantation in the world, but current most oranges and tangerines orchard all adopts homogeneous way to manage, does not consider the Spatial-Temporal Variability that orchard grows.The demand that homogeneous fertilizing management cannot meet different fruit tree individuality is carried out to orchard, causes applying fruit trees with fertilizer amount too high or too low.Therefore, obtain the nitrogenous horizontal information of every fruit tree and draw corresponding nitrogen content distributed image the quantitative fertilization tool successfully realizing every fruit tree is of great significance.
What be conventionally used to that nitrogen content detects has Kjeldahl's method and outward appearance determining method, and the former complex disposal process, cost are high, and the latter easily produces erroneous judgement due to the subjectivity of eye recognition, are all difficult to meet the demand of extensive orchard precision management at present.
Summary of the invention
For overcoming the deficiencies in the prior art, the invention provides the prediction of a kind of oranges and tangerines canopy nitrogen content based on hyperspectral technique and method for visualizing, find optimal bands combined by the correlation analysis of test figure and set up inverse model, empirically selected characteristic wave band has more science than ever mutually, can realize oranges and tangerines canopy leaves nitrogen content low cost, detect fast.
Based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique, comprise the following steps:
1) pluck covering Different Leaf Age at interior some fresh Citrus leaf sample, obtain the hyperspectral image data of each sample;
2) hyperspectral imager is utilized to obtain the Canop hyperspectrum view data of oranges and tangerines;
3) nitrogen content in each leaf sample is recorded according to Kjeldahl's method;
4) from the hyperspectral image data of each sample, the average reflectance spectra curve of each sample is extracted;
5) in the hyperspectral image data of each sample, choose different narrow band wavelength group (λ n, λ m), utilize and often organize wavelength reflected value (R corresponding in average reflectance spectra curve
λ n, R
λ m), calculate two waveband vegetation index
calculate the correlativity of TBVI value and nitrogen content again, choose correlativity the highest time corresponding TBVI value and (λ n, λ m);
6) time the highest according to all leaf sample correlativitys, corresponding TBVI value and corresponding nitrogen content, set up nitrogen content forecast model;
7) according to described Canop hyperspectrum view data, TBVI value when calculating wavelength is λ n and λ m;
8) according to step 6) in nitrogen content forecast model inverting canopy nitrogen content, and be normalized obtaining nitrogen content data, using the nitrogen content data after normalization as the gray-scale intensity of canopy each pixel elements display image, realize the visual of oranges and tangerines canopy leaves nitrogen content.
In step 3) in, each sample is dried the nitrogen content measured again to constant weight in each leaf sample.
In step 5) in, the set of wavelengths at TBVI value place corresponding when correlativity is the highest is λ n=856nm, λ m=811nm.
In step 6) in, described nitrogen content forecast model is
Y＝-102.89X+2.0058
Wherein, X be correlativity the highest time corresponding TBVI value, Y is nitrogen content.
In step 1) in, the harvesting of leaf sample should be in plant vigorous vegetative growth phase.
In step 1) and step 2) in, in the hyperspectral image data of leaf sample and Canop hyperspectrum view data, it is all 380nm to 1030nm that spectral range is, and spectral resolution is 2.8nm.
In step 2) in, what utilize hyperspectral imager to obtain is the original RGB image of oranges and tangerines canopy, and described Canop hyperspectrum view data is the original RGB image after removing background information.
Compared with prior art, the invention has the beneficial effects as follows:
(1) the present invention finds optimal bands combined by the correlation analysis of test figure and sets up inverse model, and empirically selected characteristic wave band has more science than ever mutually, can realize oranges and tangerines canopy leaves nitrogen content low cost, detect fast;
(2) distribution of oranges and tangerines canopy leaves nitrogen content is carried out visual expression directly perceived by the present invention, significant for oranges and tangerines orchard site-specific nutrient management;
(3) can according to the actual needs, utilize the method to select other vegetation index or other nutrient or other plant, set up corresponding model.
Accompanying drawing explanation
Fig. 1 is the remote sensing images schematic diagram that EO-1 hyperion instrument gathers oranges and tangerines woods;
Fig. 2 is the averaged spectrum curve of 180 Citrus leaf samples;
Fig. 3 is the averaged spectrum reflectivity of Citrus leaf at each wave band and the correlation analysis result of nitrogen content;
Fig. 4 is the two-dimensional distribution of the related coefficient of TBVI and Leaf nitrogen content;
Fig. 5 is the Citrus leaf nitrogen content calibration model of the narrow wave band TBVI based on 856nm and 811nm wavelength;
Fig. 6 is the Citrus leaf nitrogen content forecast model of the narrow wave band TBVI based on 856nm and 811nm;
Fig. 7 (a) is original RGB figure (R:660nm, G:550nm, B:460nm) of oranges and tangerines woods;
Fig. 7 (b) is for removing the oranges and tangerines woods RGB image after background information;
Fig. 7 (c) is for choosing the TBVI figure of wavelength 865nm and 811nm;
The oranges and tangerines woods nitrogen content prognostic chart that Fig. 7 (d) obtains according to TBVI figure inverting in figure (c).
Embodiment
Below in conjunction with accompanying drawing, the present invention is further described, but embodiments of the present invention are not limited thereto.
Based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique, concrete steps are as follows:
(1) Citrus leaf collection, blade and Canop hyperspectrum Image Acquisition
Select citrus reticulata "Unshiu" (CitrusunshiuMarc.) as object, this to as if originate in a Citrus Cultivars of the southeast, Asia.Pluck when being in vigor growth period in oranges and tangerines orchard and comprise tender leaf, middle period, Lao Ye totally 180 fresh blade samples.Carry out the acquisition of high spectrum image immediately in the indoor ImSpectorV10E of experiment (SpectralImagingLtd., Oulu, Finland) scanning after picking blade.In addition, in oranges and tangerines orchard, Canop hyperspectrum image is obtained with same equipment, gather schematic diagram as accompanying drawing 1, hyperspectral imager 3 is positioned at side, an oranges and tangerines trees distant place, distance to plant can photograph whole trees canopy just, correct blank 2 and be erected at tree side, computing machine 4 stores for data and analyzes.The spectral range of blade and canopy image is all 380nm to 1030nm, and spectral resolution is 2.8nm, and spatial resolution depends on the distance between sensor and testee.
After spectral reflectance data is extracted, in order to remove the noise caused by instrument, clip lower than 380nm with higher than the wavelength of 900nm.Fig. 2 shows the averaged spectrum curve of 180 Citrus leaf samples.Result shows, although all samples present similar trend in total wavelength coverage, the spectroscopic data between different sample presents very large difference.This illustrates that spectral signature can provide effective information for the nitrogen content in assessment Citrus leaf.
(2) nitrogen content of Citrus leaf is measured
After spectral scan is carried out to blade in laboratory, at once these fresh blades are stored in flash baking in the baking oven of temperatures as high 105 DEG C, again blade are positioned over after this in baking oven of 80 DEG C and slowly dry until constant weight.Subsequently, by 180 blade sample grind into powders, to measure the nitrogen content in each blade sample according to Dumas combustion with quick azotometer (ElementarAnalytical, Germany).The powder getting 50mg in each sample is measured, according to the nitrogen content in percentage calculation per unit leaf dry weight.
(3) spectroscopic data extract, Leaf nitrogen content modeling and canopy nitrogen content visual
Extract each blade sample with ENVI software and get the average reflectance spectra curve that wavelength coverage is 500-900nm (being divided into 316 wave bands), analyze the correlativity between spectroscopic data and each blade sample nitrogen content extracted.Obtain two waveband vegetation index (TBVI) by simple correlation analysis, set up the nitrogen content forecast model based on spectroscopic data on this basis.
Fig. 3 shows the averaged spectrum reflectivity of Citrus leaf at each wave band and the correlation analysis result of nitrogen content.As seen from the figure, lower than in the spectral range of 755nm, spectral reflectivity becomes positive correlation with nitrogen content, and becoming negative correlation higher than in 755nm spectral range.Wherein, the blade average reflectance spectra at 550nm and 702nm place and the related coefficient of nitrogen content the highest, be 0.6867 and 0.6861 respectively.
In recent years, two waveband vegetation index TBVI is used to replace traditional normalized differential vegetation index (NDVI) to evaluate the characteristic of various crops.This time determine can two optimal wavelengths of Accurate Prediction Citrus leaf nitrogen content for research and utilization TBVI.TBVI can by following formulae discovery:
Wherein R
λ mand R
λ nrepresentation formula medium wavelength λ n respectively, the reflected value at λ m place.Owing to comprising 316 wavelength in the hyperspectral image data in this research, the value of TBVI therefore can be calculated according to different narrow band wavelength combination of two.
Fig. 4 shows the two-dimensional distribution of the related coefficient of TBVI and Leaf nitrogen content.TBVI value calculates according to above-mentioned formula, chooses two wave band λ n (500 ~ 900nm) and λ m (500 ~ 900nm) and their spectral reflectance value is substituted into formula in 316 narrow wave bands of blade high-spectral data.Result shows, the calculating discovery TBVI value that (650 ~ 725nm) obtains within the scope of R-NIR and nitrogen content all have higher correlativity.But, the correlativity the highest (r=0.81806071) (see Fig. 5) of the TBVI value calculated by 856nm and 811nm and nitrogen content.Set up the precision of forecasting model of nitrogen content as shown in Figure 6 according to above-mentioned TBVI, the prediction of this model to whole database reaches relatively rationally precision accurately.
The visual detailed step of canopy nitrogen content is as follows:
First, choose the Citrus leaf sample spectral reflectance data that wavelength coverage is 500-900nm (being divided into 316 wave bands), and record the nitrogen content of corresponding sample with Kjeldahl's method, definition
to actual measurement nitrogen content data and any two wave band R
λ m, R
λ nthe TBVI of combination carries out correlation analysis, finds the correlativity the highest (R=0.818) of the TBVI value that 856nm and 811nm wavelength calculates and nitrogen content;
Secondly, the regression model Y=-102.89X+2.0058 (R=0.818) of actual measurement nitrogen content Y and the TBVI value X calculated with 856nm and 811nm wavelength is set up;
Finally, the TBVI image based on 865nm and 811nm wavelength is calculated according to oranges and tangerines Canop hyperspectrum image, according to the model inversion canopy nitrogen content set up in upper step, and with matlab, the nitrogen content data of each pixel elements are normalized, using the nitrogen content data after normalization as the gray-scale intensity of canopy each pixel elements display image, realize the visual of oranges and tangerines canopy leaves nitrogen content.
As shown in Figure 7, Fig. 7 a is the original RGB image of oranges and tangerines canopy, and Fig. 7 b is the oranges and tangerines canopy image after removing background information, and Fig. 7 c is the visual image of TBVI (856nm and 811nm) at canopy.Result shows, and the TBVI value of light green tender leaf is lower than the TBVI value of middle period and Lao Ye.Fig. 7 d shows the nitrogen content distribution situation of whole trees canopy, and examining image can find, the light green tender leaf with lower TBVI value demonstrates higher nitrogen content, and on the contrary, the paler colour of the middle period in figure and Lao Ye display, illustrates that nitrogen content is lower.
Claims (7)
1., based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique, it is characterized in that, comprise the following steps:
1) pluck covering Different Leaf Age at interior some fresh Citrus leaf sample, obtain the hyperspectral image data of each sample;
2) hyperspectral imager is utilized to obtain the Canop hyperspectrum view data of oranges and tangerines;
3) nitrogen content in each leaf sample is recorded according to Kjeldahl's method;
4) from the hyperspectral image data of each sample, the average reflectance spectra curve of each sample is extracted;
5) in the hyperspectral image data of each sample, choose different narrow band wavelength group (λ n, λ m), utilize and often organize wavelength reflected value (R corresponding in average reflectance spectra curve
λ n, R
λ m), calculate two waveband vegetation index
calculate the correlativity of TBVI value and nitrogen content again, choose correlativity the highest time corresponding TBVI value and (λ n, λ m);
6) time the highest according to all leaf sample correlativitys, corresponding TBVI value and corresponding nitrogen content, set up nitrogen content forecast model;
7) according to described Canop hyperspectrum view data, TBVI value when calculating wavelength is λ n and λ m;
8) according to step 6) in nitrogen content forecast model inverting canopy nitrogen content, and be normalized obtaining nitrogen content data, using the nitrogen content data after normalization as the gray-scale intensity of canopy each pixel elements display image, realize the visual of oranges and tangerines canopy leaves nitrogen content.
2., as claimed in claim 1 based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique, it is characterized in that, in step 3) in, each sample is dried the nitrogen content measured again to constant weight in each leaf sample.
3., as claimed in claim 1 based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique, it is characterized in that, in step 5) in, the set of wavelengths at TBVI value place corresponding when correlativity is the highest is λ n=856nm, λ m=811nm.
4., as claimed in claim 3 based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique, it is characterized in that, in step 6) in, described nitrogen content forecast model is
Y＝-102.89X+2.0058
Wherein, X be correlativity the highest time corresponding TBVI value, Y is nitrogen content.
5. as claimed in claim 1 based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique, it is characterized in that, the harvesting of leaf sample should be in plant vigorous vegetative growth phase.
6. as claimed in claim 5 based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique, it is characterized in that, in the hyperspectral image data of leaf sample and Canop hyperspectrum view data, spectral range is 380nm to 1030nm, and spectral resolution is 2.8nm.
7. as claimed in claim 6 based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique, in step 2) in, what utilize hyperspectral imager to obtain is the original RGB image of oranges and tangerines canopy, and described Canop hyperspectrum view data is the original RGB image after removing background information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410084466.4A CN103940748B (en) | 2014-03-10 | 2014-03-10 | Based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410084466.4A CN103940748B (en) | 2014-03-10 | 2014-03-10 | Based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103940748A CN103940748A (en) | 2014-07-23 |
| CN103940748B true CN103940748B (en) | 2016-04-06 |
Family
ID=51188510
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410084466.4A Active CN103940748B (en) | 2014-03-10 | 2014-03-10 | Based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103940748B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104614321B (en) * | 2015-01-20 | 2017-07-25 | 山东农业大学 | A kind of crop growing state method of real-time based on spectrum picture |
| CN105158177B (en) * | 2015-09-30 | 2019-03-05 | 江苏大学 | The method of hyper-spectral image technique quantitative detection solid state fermentation moisture distribution uniformity |
| CN105548455A (en) * | 2015-12-07 | 2016-05-04 | 青岛农业大学 | Method for evaluation of winter wheat blade nitrogen accumulation amount from green turning to heading |
| CN105806791B (en) * | 2016-03-22 | 2018-05-25 | 新疆农业科学院园艺作物研究所 | A kind of almond fruit puberty leaf nitrogen and phosphorus and potassium assay method |
| CN110097535B (en) * | 2019-03-22 | 2021-05-07 | 华南农业大学 | Plant leaf nitrogen content detection method and device, computer equipment and storage medium |
| CN110050555A (en) * | 2019-04-23 | 2019-07-26 | 四川农业大学 | A method of dose is judged using Citrus leaf nutrient |
| CN111652754B (en) * | 2020-05-19 | 2022-04-22 | 江西省农业科学院农业工程研究所 | Intelligent spectrum monitoring and nitrogen fertilizer quantitative diagnosis system and method for Nanfeng mandarin orange |
| CN114199790A (en) * | 2021-12-13 | 2022-03-18 | 华南农业大学 | Detection method and system for detecting nutrition status of fruit and tea trees by spectrum |
| CN115144342B (en) * | 2022-06-21 | 2024-05-28 | 河海大学 | Unmanned aerial vehicle multispectral citrus blade nitrogen content estimation method based on semi-supervised learning |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA2430737C (en) * | 2003-06-02 | 2011-12-20 | Centre De Recherche Industrielle Du Quebec | Method and apparatus for estimating surface moisture content of wood chips |
| CN100590417C (en) * | 2008-11-20 | 2010-02-17 | 北京航空航天大学 | Plant blade physiologic index non-destructive test method |
| CN102721651B (en) * | 2012-06-18 | 2014-01-15 | 浙江大学 | Detection method and system of water content in plant leaf based on multispectral image |
| CN103398957B (en) * | 2013-08-12 | 2015-11-18 | 河海大学 | The method of leaf area vertical distribution is extracted based on EO-1 hyperion and laser radar |
-
2014
- 2014-03-10 CN CN201410084466.4A patent/CN103940748B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN103940748A (en) | 2014-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103940748B (en) | Based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique | |
| Qiao et al. | UAV-based chlorophyll content estimation by evaluating vegetation index responses under different crop coverages | |
| Zhang et al. | High throughput analysis of leaf chlorophyll content in sorghum using RGB, hyperspectral, and fluorescence imaging and sensor fusion | |
| Ali et al. | Estimating leaf functional traits by inversion of PROSPECT: Assessing leaf dry matter content and specific leaf area in mixed mountainous forest | |
| Vigneau et al. | Potential of field hyperspectral imaging as a non destructive method to assess leaf nitrogen content in Wheat | |
| CN110376167A (en) | Rice leaf nitrogen content monitoring method based on unmanned plane EO-1 hyperion | |
| CN101915738B (en) | Method and device for rapidly detecting nutritional information of tea tree based on hyperspectral imaging technique | |
| Gao et al. | Improvement of chlorophyll content estimation on maize leaf by vein removal in hyperspectral image | |
| CN112557393B (en) | Wheat leaf layer nitrogen content estimation method based on hyperspectral image fusion map features | |
| CN107796764A (en) | A kind of construction method of the wheat leaf area index appraising model based on three wave band vegetation indexs | |
| CN104697943A (en) | Rape water stress nondestructive testing method | |
| CN101692037A (en) | Method for analyzing chlorophyll distribution on surface of leaves of plant by hyperspectral image and independent component | |
| CN103185695A (en) | Spectrum-based flue-cured tobacco maturity field quick judgment method | |
| CN112287886B (en) | Wheat plant nitrogen content estimation method based on hyperspectral image fusion map features | |
| CN102445421B (en) | Preparation method of leaf colour card used for rapidly diagnosing nitrogen of flue-cured tobacco | |
| CN104778349B (en) | One kind is used for rice table soil nitrogen application Classified Protection | |
| CN109870419B (en) | Method for predicting nitrogen, phosphorus and potassium content of black soil by adopting aviation hyperspectral data | |
| Peng et al. | Estimating total leaf chlorophyll content of gannan navel orange leaves using hyperspectral data based on partial least squares regression | |
| Cao et al. | Combining multispectral and hyperspectral data to estimate nitrogen status of tea plants (Camellia sinensis (L.) O. Kuntze) under field conditions | |
| CN113065230B (en) | High-spectrum inversion model for establishing rice leaf SPAD based on optimized spectral index | |
| Zarco-Tejada et al. | Remote sensing of vegetation biophysical parameters for detecting stress condition and land cover changes | |
| Li | Classification of black tea leaf water content based on hyperspectral imaging | |
| Zhang et al. | A comparative study on wheat leaf area index by different measurement methods | |
| Wang et al. | Extraction of Jujube Planting Areas in Sentinel-2 Image Based on NDVI Threshold—A case study of Ruoqiang County | |
| CN114199793B (en) | Hyperspectral-based apple leaf nitrogen content detection method |
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 |