CN103744079A - Method and system for determining planting period of sugarcane - Google Patents

Abstract

The invention is suitable for the technical field of agricultural remote sensing monitoring and provides a method for determining the planting period of a sugarcane. The method comprises obtaining sugarcane planting region multi-time phase full-polarized synthetic aperture radar data; performing registration on the obtained sugarcane planting region multi-time phase full-polarized synthetic aperture radar data; carrying out polarization interference correlation on time phase original single look complex data which is after registration; carrying out multi-look processing on the polarization interference correlation; carrying out decomposition on a polarization correlation matrix of the time phase; judging whether the plant is the sugarcane based on the data obtained from decomposition; if the plant is the sugarcane, setting the point to be a sugarcane sample point; and determining the sugarcane planting period of the sugarcane sample point based on optimum correlation coefficient and multi-time phase similarity. The method is capable of simply, rapidly, precisely and effectively distinguishing the sugarcane planting periods, and providing reliable basis for sugarcane yield estimation.

Description

A kind of definite method and system of Sugarcane Planting Stage

Technical field

The invention belongs to agricultural remote sensing monitoring technical field, relate in particular to a kind of definite method and system of Sugarcane Planting Stage.

Background technology

Crops spatial framework has reflected that mankind's agricultural production utilizes the situation of agricultural production resource in spatial dimension, is the important information of understanding crops kind, structure, distribution characteristics, is also the foundation of carrying out crop s structure adjustment and optimization.Sugarcane is important sugar crop and energy raw material, the sugarcane output of sugar of China accounts for the more than 90% of the total product of sugar, therefore, monitor sugarcane spatial framework and upgrowth situation in time,, country is formulated to sugar relevant policies, plan for the imports and exports, and sugar enterprise arranges production all significant.

Satellite remote sensing can provide the distributed intelligence of crops on room and time, than traditional manual research method, has not only saved a large amount of human and material resources, and have macroscopic view, dynamically, in time, advantage accurately.Because the phenology feature between different crops type there are differences, the remote sensing image that utilizes multidate is one of main method of crops spatial framework monitoring.The sugarcane of China 90% is distributed in south China, southwestern provinces and regions, and this region atural object distribution is broken, crop mixes seriously, and the multidate classification based on phenology feature is to distinguish sugarcane and the most effective method of other crops.

Compared with the crops such as paddy rice, wheat, corn, it is large that sugarcane has implantation time span, and diversified feature of the phase of planting is example take the Lezhou Peninsula, comprises that spring planting sugarcane, summer plantinge sugarcane, fall planting sugarcane, winter plant 5 kinds of sugarcane, stubble canes.And in sugarcane development planting process, people more and more pay attention to early, middle and late ripe, and difference is planted the reasonably combined to realize high yield and high sugar of phase, but this adds to the difficulties to the monitoring of satellite remote sensing sugarcane spatial framework simultaneously: with phase in a period of time, early, middle and late ripe, and different sugarcane remote sensing features of planting the phase there is some difference, make to apply existing multidate supervision (decision tree) classification and be difficult to realize high-precision sugarcane space distribution drawing.

South China area is cloudy rainy, and active microwave remote sensing has the sexual intercourse of not being subject to be affected, and the advantage round-the-clock, round-the-clock is monitored, becomes the significant data source that southern crops are monitored gradually.Under the prerequisite improving constantly in image resolution ratio, complete polarization has become one of important process pattern of Spaceborne SAR System of new generation.Than single (two) polarization SAR, it is more clear that complete polarization SAR portrays the architectural characteristic of target, in crops monitoring, polarization characteristic is more responsive to the biophysical parameters of reaction crop growth state, and the crop type identification based on complete polarization SAR data and spatial framework's study on monitoring have become SAR crops and monitored new developing direction.

Therefore, based on complete polarization SAR data, carrying out the detection of sugarcane physical arrangement feature, and apply on this basis multidate data simultaneously and carry out determining of Sugarcane Planting Stage, is a problem that urgency is to be studied.

Summary of the invention

The object of the present invention is to provide a kind of simple, fast, high precision, and effectively distinguish Sugarcane Planting Stage, for sugarcane the yield by estimation provides definite method and system of the Sugarcane Planting Stage of reliable basis.

The present invention is achieved in that a kind of definite method of Sugarcane Planting Stage, said method comprising the steps of:

Obtain cane-growing region multidate fully polarization synthetic aperture radar data;

The described cane-growing region multidate fully polarization synthetic aperture radar data getting is carried out to registration;

Original time phase after registration haplopia complex data is carried out to polarization interference to be concerned with;

Polarization interference is concerned with and carries out looking processing more;

To time phase polarization coherence matrix decompose;

The data that obtain according to decomposition judge whether it is sugarcane;

If judging is sugarcane, this point is made as to sugarcane sampling point;

Based on optimum coefficient of coherence and multidate similarity, judge the Sugarcane Planting Stage of described sugarcane sampling point.

Another object of the present invention is to provide a kind of Sugarcane Planting Stage fixed system really, described system comprises:

Acquisition module, for obtaining cane-growing region multidate fully polarization synthetic aperture radar data;

Registration module, for carrying out registration to the described cane-growing region multidate fully polarization synthetic aperture radar data getting;

The polarization interference module that is concerned with, relevant for original the time phase after registration haplopia complex data being carried out to polarization interference;

How depending on processing module, for carrying out looking processing to polarization interference is relevant more;

Decomposing module, for to time phase polarization coherence matrix decompose;

Judge module, judges whether it is sugarcane for the data that obtain according to decomposition;

Module is set, if be sugarcane for judging, this point is made as to sugarcane sampling point;

Sugarcane Planting Stage judge module, for judging the Sugarcane Planting Stage of described sugarcane sampling point based on optimum coefficient of coherence and multidate similarity.

In the present invention, based on multidate fully polarization synthetic aperture radar data, carry out sugarcane extraction and plant phase estimation, this scheme has been described growth conditions and the polarization characteristic of sugarcane sugarcane in tillering stage (April-May) exactly, recurrence recognition methods wherein can effectively be eliminated radar noise and crops be extracted to the drawing problem producing, and the similarity feature of phase during two of Sugarcane Planting Stage evaluation method application, describe the variation of sugarcane before and after tillering stage, and be used as with this foundation that phase of planting differentiates.The method can be used for identification, monitoring and the yield by estimation to sugarcane crops.

Accompanying drawing explanation

Fig. 1 is the realization flow schematic diagram of definite method of the Sugarcane Planting Stage that provides of the embodiment of the present invention.

Fig. 2 is the Sugarcane Planting Stage structural representation of fixed system really that the embodiment of the present invention provides.

Embodiment

In order to make object of the present invention, technical scheme and beneficial effect clearer, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein, only in order to explain the present invention, is not intended to limit the present invention.

Refer to Fig. 1, the realization flow of definite method of the Sugarcane Planting Stage providing for the embodiment of the present invention, it comprises the following steps:

In step S101, obtain cane-growing region multidate fully polarization synthetic aperture radar data;

When in embodiments of the present invention, concrete, be at the beginning of 4 months and by the end of May mutually.

In step S102, the described cane-growing region multidate fully polarization synthetic aperture radar data getting is carried out to registration;

In embodiments of the present invention, phase data are carried out registration at the beginning of 4 months and by the end of May time.

In step S103, original the time phase after registration haplopia complex data is carried out to polarization interference and be concerned with;

In embodiments of the present invention, described that original time phase after registration haplopia complex data is carried out to polarization interference is relevant, is specially:

$S_i={\left[\begin{array}{cc}{s}_{11}& s_{12}\\ s_{21}& s_{22}\end{array}\right]}_{\⇒k_i=\left[\begin{array}{ccc}{s}_{11}+s_{22}& s_{11}-s_{22}& s_{12}+s_{21}\end{array}\right]/\sqrt{2}}$

Wherein S _i, i=1,2 represents respectively at the beginning of 4 months, original haplopia complex data by the end of May, k _i, i=1,2 represent that its Pauli divides solution vector.

Within 2 o'clock, phase polarization interference is relevant, is expressed as:

$\mathrm{\Ω}=\left[\begin{array}{c}{k}_1\\ k_2\end{array}\right]\left[\begin{array}{cc}{k_1}^H& {k_2}^H\end{array}\right]=\left[\begin{array}{cc}{T}_{11}& {\mathrm{\Ω}}_{12}\\ {{\mathrm{\Ω}}_{12}}^H& T_{22}\end{array}\right]$

Wherein T ₁₁, T ₂₂the polarization coherence matrix of phase while representing respectively two, Ω ₁₂the polarization interference of phase while representing two, subscript H represents transpose conjugate.

In step S104, polarization interference is concerned with and carries out looking processing more;

In embodiments of the present invention, describedly to polarization interference is relevant, carry out looking processing more, be specially:

To polarization interference, relevant Ω carries out the processing of looking of n*m more, the value of n and m according to image orientation to distance to resolution setting, look more and process afterwards distance to consistent with azimuth resolution.Look processing as follows more:

$<\mathrm{\&Omega;}>=\frac{1}{n*m}\underset{i=1}{\overset{n*m}{\mathrm{\&Sigma;}}}\mathrm{\&Omega;}$

In step S105, to time phase polarization coherence matrix decompose;

In embodiments of the present invention, the eigenvector analysis based on coherence matrix, when application Cloude decomposes two, the polarization coherence matrix of phase decomposes, and extracts respectively its scattering entropy H and Polarization scattering angle α, odd scattering component T11.

In step S106, the data that obtain according to decomposition judge whether it is sugarcane;

In embodiments of the present invention, the whole areal coverage of image is carried out to the judgement of pointwise, if the three-dimensional polarization characteristic (H, a, T11) of phase all meets the criterion of following formula two time, think that this some vegetation of planting is sugarcane, this point is made as to sugarcane sampling point, otherwise this point is temporarily judged as non-sugarcane.

$\left\{\begin{array}{c}\mathrm{\&alpha;}\&le;40\\ H\&le;0.85\\ T_{11}\&le;\mathrm{ctg}\left(\mathrm{\&alpha;}\right)\end{array}\right.$

In step S107, if judged, be sugarcane, this point is made as to sugarcane sampling point;

As one embodiment of the present invention, after step S107, also comprise: whether the recurrence based on Neighborhood Statistics feature is further identified this sugarcane sampling point is sugarcane.

Because radar data noise is many, scattering is subject to extraneous factor, in image overlay area, still exist a large amount of actual cane-growing regions still by mistake, to be classified as non-sugarcane, therefore need to the sugarcane sampling point based on extracting carry out further sugarcane extraction above.

On the other hand, because sugarcane is subject to implantation time difference, its scattering signatures heterogeneity, divide if can produce a large amount of mistakes according to traditional supervised classification.

The embodiment of the present invention is based on neighborhood n*n statistical nature, and arranging according to the resolution of looking the rear image of processing of neighborhood n determined more.

Step 1: the sugarcane sampling point in statistics n*n neighborhood, is labeled as { ω ₁, ω ₂, Λ, ω _j, j represents sampling point number.

Step 2: sugarcane sampling point is calculated to its Wishart distance between two:

$d({\mathrm{\&omega;}}_1,{\mathrm{\&omega;}}_2)=\frac{1}{2}\{\ln \left(\right|T_1\left|\right)+\ln \left(\right|T_2\left|\right)+\mathrm{Tr}\left({T_1}^{-1}{T}_2\right)+\mathrm{Tr}\left({T_2}^{-1}{T}_1\right)\}$

Total j*(j-1) individual distance, obey one dimension Gaussian distribution, the collection of adjusting the distance carries out statistical study, calculates its average m and variances sigma.

Step 3: to the j in n*n neighborhood sugarcane sampling point, calculate its covariance matrix, as follows:

$C=\underset{i=1\mathrm{\&Lambda;j}}{\mathrm{\&Sigma;}}{T}_i,$ Be labeled as { ω }

Step 4: to the non-sugarcane point in n*n neighborhood, be labeled as

further judge one by one its attribute, rule is as follows:

If { d (η _i, ω) and≤m+3 σ, this i of mark is sugarcane, otherwise is still non-sugarcane.

Step 5: if non-sugarcane point attribute changes quantity in n*n neighborhood

return to step 1, carry out non-sugarcane point determined property next time, newly-increased sugarcane point adds in sugarcane sampling point, otherwise, finish judgement.

In step S108, based on optimum coefficient of coherence and multidate similarity, judge the Sugarcane Planting Stage of described sugarcane sampling point.

In embodiments of the present invention, step S108 is specially:

Step 10: sugarcane is sorted out to point, optimum coefficient of coherence and the similarity of phase while calculating two are as follows:

Optimum coefficient of coherence $\mathrm{\&gamma;}=\left|\mathrm{\&gamma;}\right|e^{\mathrm{i\&phi;}}=\frac{{w_1}^H{\mathrm{\&Omega;}}_{12}{w}_2}{\sqrt{{w_1}^H{T_{11}}^H{w}_1{w_2}^H{T_{22}}^H{w}_2}}$

Optimum coefficient of coherence is according to document (K.Papathanassiou and S.Cloude, " Three – stage inversion process for polarimetric sar interferometry; " IEEE Proceedings-Radar, Sonar and Navigation, vol.150, pp.125 – 134, June2003) solve.

2 o'clock phase similaritys: $\stackrel{\&OverBar;}{d}(T_{11},T_{22})=\frac{\mathrm{Tr}\left({T_{11}}^T{\left(T_{11}{T_{11}}^T\right)}^{-1}{T}_{22}\right)}{\sqrt{3}},$ In this formula, 2 o'clock phase matrixs can not be replaced position.

Step 20: Sugarcane Planting Stage judgement

If optimum coefficient of coherence | γ |>=0.6 and

being labeled as the winter plants sugarcane;

If optimum coefficient of coherence | γ |≤0.3 and

be labeled as spring planting sugarcane;

If optimum coefficient of coherence 0.6>=| γ |>=0.3 and

be labeled as stubble cane.

Step 30: adjust based on Neighborhood Statistics feature Sugarcane Planting Stage

All sugarcane sampling points in statistics n*n neighborhood,

Winter plants sugarcane point and is labeled as { ν _1i, i=1, Λ, k ₁,

Spring planting sugarcane point is labeled as { ν _2i, i=1, Λ, k ₂,

Stubble cane point is labeled as { ν _3i, i=1, Λ, k ₃.

Calculate respectively its covariance matrix:

Winter is planted sugarcane, C ₁, spring planting sugarcane C ₂, stubble cane C ₃,

To unlabelled sugarcane point in neighborhood, pointwise judges that it plants the phase, and determination methods is as follows:

Suppose sugarcane point { μ }, polarization coherence matrix T, calculates respectively itself and C ₁, C ₂, C ₃wishart distance, be labeled as d ₁, d ₂, d ₃.

If d ₁minimum, this point of mark is the winter to plant sugarcane;

If d ₂minimum, this point of mark is spring planting sugarcane;

If d ₃minimum, this point of mark is stubble cane.

Refer to Fig. 2, the Sugarcane Planting Stage structure of fixed system really providing for the embodiment of the present invention.For convenience of explanation, only show the part relevant to the embodiment of the present invention.Described Sugarcane Planting Stage really fixed system comprises:.Described Sugarcane Planting Stage really fixed system can be the unit of software unit, hardware cell or software and hardware combining.

Acquisition module 101, for obtaining cane-growing region multidate fully polarization synthetic aperture radar data;

Registration module 102, for carrying out registration to the described cane-growing region multidate fully polarization synthetic aperture radar data getting;

The polarization interference module 103 that is concerned with, relevant for original the time phase after registration haplopia complex data being carried out to polarization interference;

How depending on processing module 104, for carrying out looking processing to polarization interference is relevant more;

Decomposing module 105, for to time phase polarization coherence matrix decompose;

Judge module 106, judges whether it is sugarcane for the data that obtain according to decomposition;

Module 107 is set, if be sugarcane for judging, this point is made as to sugarcane sampling point;

Sugarcane Planting Stage judge module, for judging the Sugarcane Planting Stage of described sugarcane sampling point based on optimum coefficient of coherence and multidate similarity.

As one embodiment of the present invention, described system also comprises: identification module.

Whether identification module, for the recurrence based on Neighborhood Statistics feature, further identifying this sugarcane sampling point is sugarcane.

As another preferred embodiment of the present invention, described system also comprises:

Statistical module, for adding up the sugarcane sampling point in n*n neighborhood, is labeled as { ω ₁, ω ₂, Λ, ω _j, j represents sampling point number.

Distance calculation module, for sugarcane sampling point being calculated between two to its Wishart distance:

Covariance matrix computing module, for the j in n*n neighborhood sugarcane sampling point, calculates its covariance matrix;

Determined property module, for the non-sugarcane point in n*n neighborhood, is labeled as

further judge one by one its attribute,

Control module, if for non-sugarcane point attribute changes quantity in n*n neighborhood

return in statistical module, carry out non-sugarcane point determined property next time, newly-increased sugarcane point adds in sugarcane sampling point, otherwise, finish judgement.

In sum, the embodiment of the present invention is carried out sugarcane extraction based on multidate fully polarization synthetic aperture radar data and is planted phase estimation, this scheme has been described growth conditions and the polarization characteristic of sugarcane sugarcane in tillering stage (April-May) exactly, recurrence recognition methods wherein can effectively be eliminated radar noise and crops be extracted to the drawing problem producing, and the similarity feature of phase during two of Sugarcane Planting Stage evaluation method application, describe the variation of sugarcane before and after tillering stage, and be used as with this foundation that phase of planting differentiates.The method can be used for identification, monitoring and the yield by estimation to sugarcane crops.

One of ordinary skill in the art will appreciate that all or part of step realizing in above-described embodiment method is can carry out the hardware that instruction is relevant by program to complete, described program can be stored in a computer read/write memory medium, described storage medium, as ROM/RAM, disk, CD etc.

The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, all any modifications of doing within the spirit and principles in the present invention, be equal to and replace and improvement etc., within all should being included in protection scope of the present invention.

Claims (8)

1. a definite method for Sugarcane Planting Stage, is characterized in that, said method comprising the steps of:

Obtain cane-growing region multidate fully polarization synthetic aperture radar data;

The described cane-growing region multidate fully polarization synthetic aperture radar data getting is carried out to registration;

Original time phase after registration haplopia complex data is carried out to polarization interference to be concerned with;

Polarization interference is concerned with and carries out looking processing more;

To time phase polarization coherence matrix decompose;

The data that obtain according to decomposition judge whether it is sugarcane;

If judging is sugarcane, this point is made as to sugarcane sampling point;

Based on optimum coefficient of coherence and multidate similarity, judge the Sugarcane Planting Stage of described sugarcane sampling point.

2. the method for claim 1, is characterized in that, the described data that obtain according to decomposition judge whether it is the step of sugarcane, are specially:

In the time of two, the three-dimensional polarization characteristic of phase (H, a, T11) all meets the criterion of following formula, think that this some vegetation of planting is sugarcane, otherwise this point is temporarily judged as non-sugarcane;

。

3. the method for claim 1, is characterized in that, if be sugarcane described judging, after this point being made as to the step of sugarcane sampling point, also comprises:

Whether the recurrence based on Neighborhood Statistics feature is further identified this sugarcane sampling point is sugarcane.

4. method as claimed in claim 3, is characterized in that, whether the described recurrence based on Neighborhood Statistics feature is further identified this sugarcane sampling point is the step of sugarcane, is specially:

Sugarcane sampling point in statistics n*n neighborhood, is labeled as { ω ₁, ω ₂, Λ, ω _j, j represents sampling point number;

Sugarcane sampling point is calculated to its Wishart distance between two:

To the j in n*n neighborhood sugarcane sampling point, calculate its covariance matrix;

To the non-sugarcane point in n*n neighborhood, be labeled as further judge one by one its attribute,

Control module, if for non-sugarcane point attribute changes quantity in n*n neighborhood return in statistical module, carry out non-sugarcane point determined property next time, newly-increased sugarcane point adds in sugarcane sampling point, otherwise, finish judgement.

5. method as claimed in claim 4, is characterized in that, the described step that judges the Sugarcane Planting Stage of described sugarcane sampling point based on optimum coefficient of coherence and multidate similarity, is specially:

Sugarcane is sorted out to point, optimum coefficient of coherence and the similarity of phase while calculating two are as follows:

Optimum coefficient of coherence

2 o'clock phase similaritys:

in this formula, 2 o'clock phase matrixs can not be replaced position;

If optimum coefficient of coherence | γ |>=0.6 and

being labeled as the winter plants sugarcane;

If optimum coefficient of coherence | γ |≤0.3 and

be labeled as spring planting sugarcane;

If optimum coefficient of coherence 0.6>=| γ |>=0.3 and

be labeled as stubble cane.

6. a Sugarcane Planting Stage fixed system really, is characterized in that, described system comprises:

Acquisition module, for obtaining cane-growing region multidate fully polarization synthetic aperture radar data;

Registration module, for carrying out registration to the described cane-growing region multidate fully polarization synthetic aperture radar data getting;

The polarization interference module that is concerned with, relevant for original the time phase after registration haplopia complex data being carried out to polarization interference;

How depending on processing module, for carrying out looking processing to polarization interference is relevant more;

Decomposing module, for to time phase polarization coherence matrix decompose;

Judge module, judges whether it is sugarcane for the data that obtain according to decomposition;

Module is set, if be sugarcane for judging, this point is made as to sugarcane sampling point;

Sugarcane Planting Stage judge module, for judging the Sugarcane Planting Stage of described sugarcane sampling point based on optimum coefficient of coherence and multidate similarity.

7. system as claimed in claim 6, is characterized in that, described system also comprises:

Whether identification module, for the recurrence based on Neighborhood Statistics feature, further identifying this sugarcane sampling point is sugarcane.

8. system as claimed in claim 7, is characterized in that, described system also comprises:

Statistical module, for adding up the sugarcane sampling point in n*n neighborhood, is labeled as { ω ₁, ω ₂, Λ, ω _j, j represents sampling point number.

Distance calculation module, for sugarcane sampling point being calculated between two to its Wishart distance:

Covariance matrix computing module, for the j in n*n neighborhood sugarcane sampling point, calculates its covariance matrix;

Determined property module, for the non-sugarcane point in n*n neighborhood, is labeled as

further judge one by one its attribute,

Control module, if for non-sugarcane point attribute changes quantity in n*n neighborhood

return in statistical module, carry out non-sugarcane point determined property next time, newly-increased sugarcane point adds in sugarcane sampling point, otherwise, finish judgement.

Images