CH717361A2 - Method and system for extracting angular anomaly information from remote sensing data. - Google Patents

Abstract

The present invention discloses a method and system for extracting angular anomaly information from remote sensing data and relates to the field of mineral exploration. The method mainly comprises: storing the obtained remote sensing data in a work area in a bandwidth-wise manner to form a remote sensing data set (102), performing an angle conversion of the data in the remote sensing data set to generate a spectral angle data set (103), and extracting angle anomaly information from the remote sensing data the spectral angle data set using an orthogonal decomposition technology (104). The present invention combines spectral features and spectral differences and extracts angular anomaly information, which provides better results than extracting ordinary anomaly information from remote sensing data.

Description

TECHNICAL PART

The present invention relates to the field of mineral exploration and, more particularly, to a method and system for extracting information about angular anomalies from remote sensing data.

BACKGROUND

In practice it has been shown that the extraction of information about common anomalies from remote sensing data plays an important role in mineral exploration in the desert region of the Gobi. However, the extracted information about common anomalies contains much interference information and unidentifiable information. This is because the extraction of common anomalies information from remote sensing data focuses only on the spectral characteristics of the data, not on the spectral angle information.

SUMMARY

The present invention provides a method and system for extracting angular anomaly information from remote sensing data that combines spectral features and spectral differences and extracting angular anomaly information, which provides better results than extracting ordinary anomaly information from remote sensing data .

In order to achieve the above object, the present invention offers the following solutions: A method for extracting angular anomaly information from remote sensing data includes: obtaining multiband remote sensing data in a work area; Storage of the remote sensing data per band to form a remote sensing data set, the remote sensing data of each band in the work area being composed of a two-dimensional matrix with coordinates; Implementation of the angle conversion of the data in the remote sensing data set to generate a spectral angle data set; and extracting angle anomaly information from the remote sensing data from the spectral angle data set using an orthogonal decomposition technique.

Optionally, after the extraction of angle anomaly information from the remote sensing data from the spectral angle data set using an orthogonal decomposition technology, the method includes a further overlay of the angle anomaly information from the remote sensing data with the multiband remote sensing data to produce an image of the remote sensing Output target position information.

A system for extracting angular anomaly information from remote sensing data includes: a module for obtaining remote sensing data in a work area, configured to receive multiband remote sensing data in a work area; a remote sensing data set generation module configured to store the remote sensing data band by band to form a remote sensing data set, the remote sensing data of each band in the work area being composed of a two-dimensional matrix of coordinates; a module for generating a spectral angle data set configured to perform an angular conversion of the data in the remote sensing data set to generate a spectral angle data set; and an angular anomaly information extraction module configured to extract angular anomaly information from the remote sensing data from the spectral angular data set using an orthogonal decomposition technology.

Optionally, the system includes further options: an image output module configured to overlay the angular anomaly information from the remote sensing data with the multiband remote sensing data to output an image of the remote sensing target position information.

According to examples provided by the present invention, this invention discloses the following technical effects.

The present invention provides a method and system for extracting angular anomaly information from remote sensing data. The method mainly comprises: storing the obtained remote sensing data in a work area in a bandwidth-wise manner to form a remote sensing data set; Performing an angular conversion of the data in the remote sensing data set to generate a spectral angle data set; and extracting angle anomaly information from the remote sensing data from the spectral angle data set using an orthogonal decomposition technology. The present invention combines spectral features and spectral differences and extracts angular anomaly information, which provides better results than extracting ordinary anomaly information from remote sensing data.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to describe the technical solutions in the examples of the present invention or in the prior art in more detail, the accompanying drawings required for the examples are briefly described below. The drawings in the following description show only a few examples of the present invention and one skilled in the art can derive still other examples of the invention from these accompanying drawings without creativity.

FIG. 1 is a flow diagram of a method for extracting angular anomaly information from remote sensing data in accordance with the present invention.

FIG. 2 is a frequency domain histogram of angles corresponding to an angular anomaly information extraction region in accordance with the present invention.

FIG 3 is a schematic representation of the positions of the angle anomaly information in accordance with the present invention.

FIG 4 is a diagram of remote sensing target position information in accordance with the present invention.

FIG 5 is a structural diagram of a system for extracting angular anomaly information from remote sensing data in accordance with the present invention.

DETAILED DESCRIPTION

In the following, the technical solutions in the examples of the present invention will be clearly and completely described, reference being made to accompanying drawings in the examples of the present invention. Obviously, the examples described are only a part and not all of the examples of the present invention. All other examples obtained by one skilled in the art on the basis of the examples of the present invention without any creative effort also fall within the scope of the present invention.

The present invention provides a method and system for extracting angular anomaly information from remote sensing data that combines spectral features and spectral differences and extracts angular anomaly information, which provides better results than extraction of general anomaly information Remote sensing data.

In order to make the above-mentioned objects, features and advantages of the present invention better understood, the present invention is described in more detail below with reference to the accompanying drawings and detailed examples.

An "angle anomaly" is an anomaly that is extracted using an anomaly extraction technique based on a spectral angle data set that can overcome the errors caused by the usual anomaly extraction of data values and that take advantage of the geometric anomaly extraction and statistical classification combined to accurately identify anomalies. It consists of three steps: angle construction, angle anomaly extraction, and anomaly slicing and optimization.

As shown in FIG 1, a method for extracting angular anomaly information from remote sensing data provided by the present invention comprises the following steps:

Step 101: Acquisition of multiband remote sensing data in a work area.

The received multiband remote sensing data of the working area (mining area) can be hyperspectral data, multispectral data and other optical remote sensing data such as ASTER and LANDSAT.

Step 102: storing the remote sensing data band by band to form a remote sensing data set, the remote sensing data of each band in the work area being composed of a two-dimensional matrix with coordinates.

The remote sensing data obtained are stored (output) band by band. For example, BAND1 represents a first band and BAND2 represents a second band. There are M × N bands in total. Each band consists of a two-dimensional matrix with coordinates. The generated remote sensing data set is shown in the following formula.

Step 103: Performing an angle conversion on the data in the remote sensing data set in order to generate a spectral angle data set. Specifically, calculate an average of the remote sensing data in the remote sensing dataset, i.e.:

BANDI represents Band I two-dimensional matrix data (i.e., Band I remote sensing data). Use the mean value as the origin to perform an angular conversion of the remote sensing dataset to obtain angular data, i.e.:

Bi represents a specific value in two-dimensional matrix data of the band i, and ϕj represents spectral angle data obtained after the data conversion of the band j. A newly generated spectral angle data set is as follows:

[0028] Φ1 represents spectral angle data obtained after Volume 1 data conversion. Alternatively, the angular conversion can be carried out in one band. For example, the angle conversion in band BANDI can be as follows:

Bi represents a specific value in the two-dimensional matrix data of the band i, Bijk represents a specific value in row j and column k of the two-dimensional matrix data of the band I, and θi represents spectral angle data obtained after the band data conversion. A newly generated spectral angle data set of the band looks like this:

[0030] Θ1 represents spectral angle data obtained after converting the remote sensing data of the first band.

Step 104: Extraction using an orthogonal decomposition technique of angle anomaly information from the remote sensing data from the spectral angle data set.

(1) Use of a skew coefficient and a kurtosis coefficient to select an area shown in FIG 3 from the spectral angle data set for the extraction of angle anomaly information.

Assumption of an area p × q in the spectral angle data sets ANGSET and ACRSET. A band pixel Θst (s = 1, ..., p; t = 1, ..., q) of a remote sensing image in this area lies within [x0, xp], is a mean value of the pixels Θ and a standard deviation is σ. Using the Skew coefficient and the kurtosis coefficient to determine a histogram. FIG 2 is a histogram of the angles in the frequency range that correspond to the range in which the information about the angle anomaly was extracted.

The skew coefficient corresponds to the following formula: ε1 is a given small positive number.

The kurtosis coefficient corresponds to the following formula: ε2 is a given small positive number.

Then this p × q area is the selected data X.

(2) Using the orthogonal decomposition algorithm to process the data in the area of extracting the angle anomaly information to extract the angle anomaly information from the remote sensing data.

Using the orthogonal decomposition algorithm for the selected data X to extract the angular anomaly information from the remote sensing data.

The principle is as follows: First, the origin of the coordinates is shifted so that the mean value is zero. Then the coordinates are rotated so that a coordinate axis coincides with a direction in which most of the data is distributed. The rotated axis is a first orthogonal base that makes up the largest portion of the total information. A coordinate axis perpendicular to this rotated axis represents a direction of the remaining information, which represents a second orthogonal base. In a multidimensional space with more than two dimensions, this processing continues until a set of rectangular coordinate axes is determined. All information is distributed (consumed) on these right-angled coordinate axes. The information cannot all be contained in a secondary orthogonal base. Instead, the number of orthogonal bases is equal to the number of original parameters. The total amount of information of all orthogonal bases is equal to that before the conversion, i.e. the preservation of the information.

The data of the bands are mapped onto new orthogonal bases, the amount of which matches that of the bands. Every orthogonal basis is formed by the linear addition of eigenvectors. Mathematics is about finding some new variables ξ1, ξ2, ..., ξu that are linear functions of X but are not related to each other, that is:

Actually, the point is to find the u2 constants Lfg (f, g = 1, ..., u) expressed as a matrix:

In the formula, L is an intrinsic matrix, each Lfg is a component of the eigenvector and λ is an eigenvalue of a matrix C. λ and L have the following characteristics: is called a trace, or the total amount of information.

The values of L (i.e. the principal components) corresponding to the various values of λ are linearly uncorrelated.

It is known from linear algebra that the eigenpolynomial of the covariance matrix is C det (λI-C), and the root λ of this eigenpolynomial is the eigenvalue of the covariance matrix C.

The calculation is as follows: Covariance matrix C:

Eigenvalue λ:

Eigenvector L: (λI-C) L = 0.

When a coordinate axis of data in band N is transposed, the covariance matrix is also transformed. After the transformation, the covariance between the bands becomes zero.

The sum of the squares of the distances between all points and the centroid is the sum of the eigenvalues, and this sum can be expressed as S. In a sense, a ratio of the information amount of the first orthogonal base to the total information amount is λ1 / S, a ratio of the information amount of the first two orthogonal bases to the total information amount is (λ1 + λ2) / S, and so on. For example, one can say that “the first four components make up u% of the amount of information”.

An orthogonal base eigenvalue is a mean square error value that is introduced into a corresponding eigenvector when the orthogonal base is removed.

The eigenvectors obtained are matched with each band involved in the orthogonal transformation to select an eigenvector that satisfies the characteristics of the change anomalies, which is usually a fourth vector. Table 1 shows the correspondence.

When the characteristic of an anomaly is Va4> Vb4 <Vc4> Vd4, Va4 and Vc4 must have opposite signs to Vb4 and Vd4, while Va4 and Vc4 and Vb4 and Vd4 must have the same sign. The eigenvector 4 used for anomaly slicing requires that Vc4 be a positive value. If it is a negative value, it must be converted to a positive value using the formula below.

Vc4T is a result obtained after negative sign conversion of Vc4.

(3) Optimization and cutting of the angle anomaly information from the remote sensing data on the basis of a normal distribution principle in order to obtain the final angle anomaly information from the remote sensing data shown in FIG 3.

Performing pre-orthogonal decomposition and transformation processing to implement the normal distribution for the histogram of each band as well as a histogram of the principal components of the transformed anomaly (namely, an eigenvector). Then the anomaly information is sliced according to the principle of normal distribution. A normal distribution formula is as follows:

X stands for a random variable and σ stands for a standard error. For the orthogonal transformation, σ is called the standard deviation and is calculated as follows:

N is a set of samples, x is a mean value, and xi is a value of each sample. σ can be used to represent a scale of a normal distribution curve in anomaly slicing or data slicing. For example, after the orthogonal transformation, the mean value x can be understood as the area background, and (X + kσ) is used to determine a lower limit and a degree of intensity of the anomaly. In general, ± 4σ is used as the upper and lower limit.

This makes the anomaly slicing more objective. The anomaly level is calculated using the following formulas: L = 127.5 + k * SF or L = 127.5 + k * 127.5 / 4; H = L + 1.

H and L represent the upper limit and the lower limit of slicing; k is a multiple; σ is a standard deviation; and SF is a scale factor, where σ and SF are given in the orthogonal transformation report.

Step 105: Overlaying the angle anomaly information from the remote sensing data with the multi-band remote sensing data to output an image of the remote sensing target position information.

The information about the angle anomaly is superimposed on an original image in order to output a superimposed image that is suitable for eye observation, e.g. the information image shown in FIG. 4 about the target position of remote sensing. The final image in JPG or TIF format is output via software.

As shown in FIG 5, the present invention also provides a system for extracting information about angular anomalies from remote sensing data, including a workspace remote sensing data collection module 201 configured to collect multiband remote sensing data in a workspace; a remote sensing record generation module 202 configured to store the remote sensing data band by band to form a remote sensing data set, the remote sensing data of each band in the work area being composed of a two-dimensional matrix of coordinates; a spectral angle data set generation module 203 configured to angularly convert the data in the remote sensing data set to generate a spectral angle data set; an angle anomaly information extraction module 204 configured to extract angle anomaly information from the remote sensing data from the spectral angle data set using an orthogonal decomposition technology; and an image output module 205 configured to overlay the angular anomaly information from the remote sensing data with the multiband remote sensing data to output an image of the remote sensing target position information.

The module 203 for generating spectral angle data sets comprises in particular an average calculation unit configured to calculate an average of the remote sensing data in the remote sensing data set; and a spectral angle data set generation unit configured to use the mean value as the origin to perform angle conversion of the data in the remote sensing data set to obtain angle data and generate the spectral angle data set.

The angular anomaly information extraction module 204 includes in particular an extraction area determining unit configured to use a skew coefficient and a kurtosis coefficient to select an extraction area for angle anomaly information from the spectral angle data set; an angle anomaly information extraction unit configured to use the orthogonal decomposition algorithm to process data in the angle anomaly information extraction area to extract the angle anomaly information from the remote sensing data; and an angular anomaly information optimization unit configured to optimize and intersect the angular anomaly information from the remote sensing data to obtain final angular anomaly information from the remote sensing data.

The present invention discloses a method and system for extracting information about angular anomalies from remote sensing data to enable a new technical approach to extracting information about potential, hidden and weak anomalies. According to the present invention, a spectral angle data set is created and the angle anomaly information is extracted from the spectral angle data set using an orthogonal decomposition technology to determine the remote sensing target position information. This can solve problems such as difficult extraction of hidden, potential and weak anomaly target position information, many false anomalies and inaccurate extracted results, and purposefully improve the anomaly information, thereby providing technical assistance for accurately locating targets. The present invention promotes mineral exploration with less time, manpower and material resources. It is a new technique that can promote production and development. The present invention has been applied to mineral exploration and has successfully discovered several mining sites in China's flat, arid, and vegetation-covered areas.

Each example of the present specification is described progressively, each example focuses on the difference from other examples, and like and similar parts between the examples may be related to each other. For a system disclosed in the examples, since it corresponds to the method disclosed in the examples, the description is relatively simple and reference can be made to the description of the method.

Several examples are used throughout this paper to illustrate the principles and implementations of the present invention. The description of the foregoing examples serves to illustrate the method of the present invention and its basic principles. Moreover, those of ordinary skill in the art can make various modifications relating to specific implementations and areas of application in accordance with the teachings of the present invention. In conclusion, the content of the present specification is not to be construed as a limitation on the present invention.

Claims (8)

1. A method for extracting angular anomaly information from remote sensing data, comprising: Acquisition of multiband remote sensing data in a workspace; Storing the remote sensing data per band to form a remote sensing data set, the remote sensing data of each band in the work area being composed of a two-dimensional matrix of coordinates; Implementation of the angle conversion of the data in the remote sensing data set to generate a spectral angle data set; and Extracting angular anomaly information from the remote sensing data from the spectral angular data set using an orthogonal decomposition technique. 2. The method of extracting angle anomaly information from remote sensing data according to claim 1, wherein after extracting angle anomaly information from the remote sensing data from the spectral angle data set using an orthogonal decomposition technology, the method further comprises: Overlaying the angular anomaly information from the remote sensing data with the multiband remote sensing data to output an image of the remote sensing target position information. 3. The method for extracting angular anomaly information from remote sensing data according to claim 1, wherein performing the angular conversion of the data in the remote sensing data set to generate a spectral angle data set specifically comprises: Calculating an average value of the remote sensing data in the remote sensing data set; and using the mean as the origin to angularly convert the data in the remote sensing dataset to obtain angular data and generate the spectral angular dataset. 4. The method of extracting angular anomaly information from remote sensing data according to claim 1, wherein extracting angular anomaly information from the remote sensing data from the spectral angle data set using an orthogonal decomposition technology specifically comprises: Using a skew coefficient and a kurtosis coefficient to select an information extraction range for angle anomalies from the spectral angle data set; Using the orthogonal decomposition algorithm to process data in the field of extracting angular anomaly information to extract the angular anomaly information from the remote sensing data; and Optimization and slicing of the angular anomaly information from the remote sensing data based on a normal distribution principle to obtain final angular anomaly information from the remote sensing data. 5. A system for extracting angular anomaly information from remote sensing data, comprising a module for obtaining remote sensing data in a work area, configured to receive multiband remote sensing data in a work area; a remote sensing data set generation module configured to store the remote sensing data band by band to form a remote sensing data set, the remote sensing data of each band in the work area being composed of a two-dimensional matrix of coordinates; a module for generating a spectral angle data set configured to perform an angular conversion of the data in the remote sensing data set to generate a spectral angle data set; and an angular anomaly information extraction module configured to extract angular anomaly information from the remote sensing data from the spectral angular data set using an orthogonal decomposition technology. 6. The system for extracting angular anomaly information from remote sensing data according to claim 5, further comprising: an image output module configured to overlay the angular anomaly information from the remote sensing data with the multiband remote sensing data to output an image of the remote sensing target position information. 7. The system for extracting angular anomaly information from remote sensing data according to claim 5, wherein the module for generating a spectral angular data set specifically comprises: an average calculation unit configured to calculate an average of the remote sensing data in the remote sensing data set; and a spectral angle data set generation unit configured to use the mean value as the origin to perform angle conversion of the data in the remote sensing data set to obtain angle data and generate the spectral angle data set. 8. The system for extracting angular anomaly information from remote sensing data according to claim 5, wherein the module for extracting angular anomaly information specifically comprises: an extraction area determining unit configured to use a skew coefficient and a kurtosis coefficient to select an extraction area for angle anomaly information from the spectral angle data set; an angle anomaly information extraction unit configured to use the orthogonal decomposition algorithm to process data in the angle anomaly information extraction area to extract the angle anomaly information from the remote sensing data; and an angular anomaly information optimization unit configured to optimize and intersect the angular anomaly information from the remote sensing data to obtain final angular anomaly information from the remote sensing data.