CN108304766A - A method of based on high-definition remote sensing screening dangerous material stockyard - Google Patents
A method of based on high-definition remote sensing screening dangerous material stockyard Download PDFInfo
- Publication number
- CN108304766A CN108304766A CN201711312959.9A CN201711312959A CN108304766A CN 108304766 A CN108304766 A CN 108304766A CN 201711312959 A CN201711312959 A CN 201711312959A CN 108304766 A CN108304766 A CN 108304766A
- Authority
- CN
- China
- Prior art keywords
- remote sensing
- scale
- image
- point
- dangerous material
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
- G06V10/449—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/757—Matching configurations of points or features
Abstract
The invention discloses a kind of methods based on high-definition remote sensing screening dangerous material stockyard, include the following steps：It is distinguished first by gray level co-occurrence matrixes, and target area is extracted according to the uniformity coefficient of gray scale；Secondly, it is used as the invariant features point of interest response extraction image local maximum and minimum point as image by carrying out convolution to scale parameter and directioin parameter in target area；Furthermore the matching of model and image is carried out by Hausdorff distances, and translation transformation is carried out by invariant features point and eliminates space error, matching is fitted to invariant features point；It finally passes sequentially through contrast characteristic's point to screen, obtains dangerous material stockyard；The substantive characteristics of remote sensing images is utilized, effective, reliable matching result can be generated, there is preferable fault-tolerant ability, situation that can be shielded to target part in image and image comprising multiple targets simultaneously is handled, it is suitable for the actual conditions in dangerous material stockyard, screening recognition efficiency is high.
Description
Technical field
It is specially a kind of that dangerous material stockyard is screened based on high-definition remote sensing the present invention relates to remote sensing recognition technical field
Method.
Background technology
In traditional, compared with low resolution remote sensing image data, high-definition picture has following high-definition picture
Several features：First, high-resolution remote sensing image is influenced by aerial image technology, and actual image data is for the light that utilizes
It composes wave band to reduce, and with the reduction of spectral information, same object different images in image, foreign matter the phenomenon that composing with will largely exist, difference
Ground object target is difficult to realize statistics extraction in image spectral domain；Secondly, with the raising of image resolution, in the same region
Interior, the information obtaining ability of atural object will significantly increase, and at the same time, atural object classification will be more various in image, numerous
Atural object in object is identified out is highly difficult；Finally, due to the atural object texture and shape that show in high resolution image
Shape information is more obvious, in the past based in, the Spectral unmixing method of the remote sensing image of low resolution is difficult to be directly applied to high score
It distinguishes in remote sensing image data interpretation.
And it is carried out by high-resolution remote sensing image in the method in dangerous material stockyard in identification harbour in remote sensing images
Identification will be more difficult, this is because not having special spectrum feature and shape feature in dangerous material stockyard, can not establish effectively
Identification database, therefore using traditional remote sensing images method computer it is effective identification in be hard to work.
Invention content
In order to overcome the shortcomings of that prior art, the present invention provide a kind of based on high-definition remote sensing screening dangerous material heap
The method of field, is utilized the substantive characteristics of remote sensing images, can also be generated while having evaded correction calculation effective, reliable
Matching result has preferable fault-tolerant ability, can be hidden simultaneously to target part in image and image comprising multiple targets
The case where covering is handled, and is suitable for the actual conditions in dangerous material stockyard, screening recognition efficiency is high, can effectively solve the problem that background skill
The problem of art proposes.
The technical solution adopted by the present invention to solve the technical problems is：One kind screening dangerous material based on high-definition remote sensing
The method in stockyard, includes the following steps：
S100, target area is obtained based on textural characteristics, different textural characteristics and knot is distinguished by gray level co-occurrence matrixes
Structure characteristic, and target area is extracted according to the uniformity coefficient of gray scale；
S200, feature point extraction is carried out by the method for scale-space, by scale parameter and side in target area
Convolution, which is carried out, to parameter responds the invariant features point of extraction image local maximum and minimum point as image as interest；
S300, by translation vector eliminate invariant features point space error, by Hausdorff distance carry out model and
The matching of image, and translation transformation is carried out by invariant features point and eliminates space error, matching is fitted to invariant features point；
S400, invariant features point library is established, remote sensing features point is screened one by one, matched inconvenient characteristic point will be fitted
Sequence composition invariant features point library successively, and pass sequentially through contrast characteristic's point and screen, obtain dangerous material stockyard.
As a kind of preferred technical solution of the present invention, gray level co-occurrence matrixes distinguish different textural characteristics and structure feature
Specific steps：
Setting the gray level that L indicates image local area, then gray level co-occurrence matrixes are expressed as the matrix of LxL in the region,
That is Pθ, d=[PI, j]LxL, wherein matrix element PijIt is θ to be expressed as direction in image local area, and distance is d, gray scale
Value is respectively frequency of the pixel to appearance of i and j；
Co-occurrence matrix feature is extracted by statistical measurement by the frequency of appearance, obtains the grey scale change and line in the region
Manage feature.
As a kind of preferred technical solution of the present invention, spatial frequency and location information are being extracted by gray level co-occurrence matrixes
It filters progress textural characteristics by Gabor afterwards finely to extract, specific method is：
S101, set the functional form of Gabor filter as Wherein x, y are position coordinates, and j is space domain characteristic value, and w is concussion frequency, σxAnd σyFiltering is indicated respectively
Standard deviation of the device in the directions x and the directions y；
S102, scale parameter u and directioin parameter v is added, then filter function is：
guv(x, y)=a-uG (x ', y '), wherein u, v are integer, a ＞ 1,
And derivative
DerivativeWherein K is total direction number；
S103, m scale is carried out by filter function input image I (x, y), the Gabor in n direction is converted, and right
The feature of unified scale different directions is averaged to obtain m textural characteristics subband to get the characteristic vector tieed up to a m as shadow
Texture feature information X={ the x of each pixel (x, y) as in1, x2..., xm}。
As a kind of preferred technical solution of the present invention, in step s 200, the specific algorithm of scale-space method is：
S201, the standard deviation Gauss equation for corresponding to different scale is set as L (u, v, σ)=G (u, v, σ) * I (u, v), wherein
U is scale parameter, and v is directioin parameter, and σ is scale, and G (u, v, σ) is two-dimensional Gaussian function, and I (u, v) is the ruler of remote sensing images
Degree-directivity function；
S202, two-dimensional Gaussian function is calculated
S203, scale-directivity function I (u, the v)=L (u, v, σ) for setting remote sensing images carry out cycle calculations, until scale σ
Equal to parameter value, and the reference value of scale σ is 1.6.【Technique study P19】
It should be carried out two-way as a kind of preferred technical solution of the present invention when carrying out the calculating of scale-space method
It calculates.
As a kind of preferred technical solution of the present invention, the Hausdor distances are specifically defined as：Set two remote sensing
Characteristic point A={ a1, a2..., apAnd B={ b1, b2..., bq, then Hausdor distance be H (A, B)=max (h (A, B), h (B,
A)), wherein| | | | it is distance measure.
As a kind of preferred technical solution of the present invention, in Hausdor distances calculate, acceleration strategy, which may be used, to be had
It is calculated in the range of limit.
As a kind of preferred technical solution of the present invention, in step S300, invariant features point carries out the tool of translation transformation
Body step is：
Two S301, setting feature point set A={ a1, a2..., apAnd B={ b1, b2..., bq, and feature point set is taken to make
For the reference set of range conversion；
S302, settingIt is characterized the distance map of point set A, sets D 'It is characterized the distance map of point set B, then there are asked relative to each other when translation transformation between A and B
Take Hausdor distances；
S303, the ratio of defined feature point set A and feature point set B or the ratio of feature point set B and feature point set A are
Threshold value is converted, and it is F (x, y) to set the distances of the opposite Hausdor between A and B, during carrying out translation transformation so that F
It is effective translation transformation that the change threshold that (x, y) is minimized, which is less than 1,.
As a kind of preferred technical solution of the present invention, after composition invariant features point library, specific screening step is such as
Shown in lower：
S401, setting time coefficient are γ1, cloud amount coefficient is γ2, resolution rate coefficient is γ3, wherein 0≤γ≤1, and
γ1+γ2+γ3=1；
S402, remote sensing image weighted value be γ=γ1xλ1+γ2xλ2+γ3xλ3, wherein λ1、λ2And λ3It is respectively corresponding
The weight of coefficient；
S403, the weighted value that the record per data is calculated according to the weighted value of above-mentioned remote sensing image, and to all remote sensing
Data are screened, and the maximum data of weighted value are charged to filter record collection.
As a kind of preferred technical solution of the present invention, duplicate removal processing is carried out to the filter record collection of acquisition, is obtained new
Result set is retrieved and judges data that filter record is concentrated whether in result set, there is no result set is then added, until completing
The retrieval of all data.
Compared with prior art, the beneficial effects of the invention are as follows：Present invention utilizes the substantive characteristics of remote sensing images, can
Overcome image to change the influence brought in imaging process conditional, the application of correcting algorithm can be overcome, evades the fortune of correction
It calculates, effective, reliable matching result can also be generated while having evaded correction calculation, to smaller in high-definition picture
Labour, which influences insensitive that is, smaller disturbance, not to be influenced to carry out in screening process, this is just effectively directly neglected on source
Data are slightly interfered, there is preferable fault-tolerant ability, it can on the basis of retaining legacy data characteristic by the comparison of characteristic point
It is handled with situation shielded to target part in image and image comprising multiple targets simultaneously, is suitable for dangerous material heap
Actual conditions, not complicated data calculation process, whole process is simple and direct, and it is conventional screening calculation amount that calculating process, which takes,
30%-50%, screening recognition efficiency be well improved.
Description of the drawings
Fig. 1 is the overall structure diagram of the present invention.
Specific implementation mode
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation describes, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Embodiment：
As shown in Figure 1, the present invention provides a kind of methods based on high-definition remote sensing screening dangerous material stockyard, including such as
Lower step：
Step S100, textural characteristics are based on and obtain target area, different textural characteristics are distinguished by gray level co-occurrence matrixes
And architectural characteristic, and target area is extracted according to the uniformity coefficient of gray scale.
Step S101, setting L indicates the gray level of image local area, then gray level co-occurrence matrixes are expressed as in the region
The matrix of LxL,
That is Pθ, d=[PI, j]LxL, wherein matrix element PijIt is θ to be expressed as direction in image local area, and distance is d, gray scale
Value is respectively frequency of the pixel to appearance of i and j；
Co-occurrence matrix feature is extracted by statistical measurement by the frequency of appearance, obtains the grey scale change and line in the region
Manage feature.For example, if intensity profile is uneven in region, the value of the main focusing element of matrix can be bigger.
In the present embodiment, it is also necessary to attention, due in the processing procedure of gray level co-occurrence matrixes, the ash in region
Degree is random distribution, and similar frequency will be presented in the element in co-occurrence matrix, and this frequency will in high-definition picture
Will produce it is serious obscure interference, need to carry out fine processing, be convenient for identification and utilization of the subsequent step to textural characteristics, logical
It is finely extracted, is had by Gabor filtering progress textural characteristics after crossing gray level co-occurrence matrixes extraction spatial frequency and location information
The method of body is：
Step S101, set the functional form of Gabor filter as Wherein x, y are position coordinates, and j is space domain characteristic value, and w is concussion frequency, σxAnd σyRespectively
Standard deviation of the expression filter in the directions x and the directions y；
Step S102, scale parameter u and directioin parameter v is added, then filter function is：
guv(x, y)=a-uG (x ', y '), wherein u, v are integer, a ＞ 1,
And derivative
DerivativeWherein K is total direction number；
Step S103, by filter function input image I (x, y) carry out m scale, the Gabor transformation in n direction,
And it is averaged to obtain m textural characteristics subband to the feature of unified scale different directions to get the characteristic vector work tieed up to a m
For the texture feature information X={ x of each pixel (x, y) in image1, x2..., xm}。
Step S200, feature point extraction is carried out by the method for scale-space, by scale parameter in target area
Convolution, which is carried out, with directioin parameter responds the invariant features of extraction image local maximum and minimum point as image as interest
Point.
The extraction of feature based point is due to detecting characteristic point in the target area, having operational efficiency high, extract feature
The advantages that abundant, convenient for the application in high-resolution remote sensing image.
In step s 200, the specific algorithm of scale-space method is：
Step S201, set the standard deviation Gauss equation of corresponding different scale as L (u, v, σ)=G (u, v, σ) * I (u,
V), wherein u is scale parameter, and v is directioin parameter, and σ is scale, and G (u, v, σ) is two-dimensional Gaussian function, and I (u, v) is remote sensing figure
Scale-directivity function of picture；
Step S202, two-dimensional Gaussian function is calculated
Step S203, scale-directivity function I (u, the v)=L (u, v, σ) for setting remote sensing images carries out cycle calculations, until ruler
It spends σ and is equal to parameter value, and the reference value of scale σ is 1.6.
Step S300, the space error that invariant features point is eliminated by translation vector is carried out by Hausdorff distances
The matching of model and image, and translation transformation is carried out by invariant features point and eliminates space error, invariant features point is intended
Close matching.
The Hausdor distances are specifically defined as：Set two remote sensing features point A={ a1, a2..., apAnd B=
{b1, b2..., bq, then Hausdor distances are H (A, B)=max (h (A, B), h (B, A)), wherein| | | | it is distance measure.
By above-mentioned definition it is found that Hausdor distances have weighed the whole matching degree of point set A and point set B, it is not suitable for
Therefore the subset of the subset and point set B of point set A when carrying out Hausdor distance calculating, that is, is carrying out scale-sky
Between method calculating when, two-way calculating should be carried out, using two-way Hausdor distances overcome subset cannot participate in calculate lack
It falls into, realizes comprehensive covering, the consistency matching of subset can be realized while point set is matched.
And among the above, the matching of model and image is carried out using the Hausdor distances of two-way part, this is the spy by image
What point determined, because image generally comprises multiple targets, and target in the picture is there may be the shielded situation in part, into
The two-way calculating of row can carry out the feature points of matched image section and model part by being arranged, and then overcome respectively
The above problem.
In step S300, invariant features point carry out translation transformation the specific steps are：
Step S301, two feature point set A={ a are set1, a2..., apAnd B={ b1, b2..., bq, and take characteristic point
Collect the reference set as range conversion；
Step S302, it setsIt is characterized the distance map of point set A, sets D 'It is characterized the distance map of point set B, then there are asked relative to each other when translation transformation between A and B
Take Hausdor distances；
Step S303, the ratio of the ratio of defined feature point set A and feature point set B or feature point set B and feature point set A
To convert threshold value, and sets the distances of the opposite Hausdor between A and B and make during carrying out translation transformation for F (x, y)
It is effective translation transformation to obtain the change threshold that F (x, y) is minimized to be less than 1.
During carrying out translation transformation, by above-mentioned it is found that not all above-mentioned transformation is meaningful,
Therefore it can be calculated in the range that translation collects very limited by acceleration strategy.
Step S400, invariant features point library is established, remote sensing features point is screened one by one, will be fitted matched inconvenient special
Sequence forms invariant features point library to sign point successively, and passes sequentially through contrast characteristic's point and screen, and obtains dangerous material stockyard.
After composition invariant features point library, specific screening step is as follows：
Step S401, setting time coefficient is γ1, cloud amount coefficient is γ2, resolution rate coefficient is γ3, wherein 0≤γ≤
1, and γ1+γ2+γ3=1；
Step S402, the weighted value of remote sensing image is γ=γ1xλ1+γ2xλ2+γ3xλ3, wherein λ1、λ2And λ3Respectively
The weight of coefficient of correspondence；
Step S403, the weighted value of the record per data is calculated according to the weighted value of above-mentioned remote sensing image, and to all
Remotely-sensed data is screened, and the maximum data of weighted value are charged to filter record collection.
In above-mentioned screening step, it is necessary first to according to the condition of retrieval, such as resolving range, geographic area and
Time range etc. retrieves satisfactory all remote sensing image data record sets, and weight system is determined further according to actual demand
Number carries out above-mentioned steps.
Duplicate removal processing is carried out to the filter record collection of acquisition, obtains new result set, retrieve and judges that filter record is concentrated
Data whether in result set, there is no result set is then added, until completing the retrieval of all data.
In conclusion the main characteristic of the invention lies in that：
(1) substantive characteristics of remote sensing images is utilized, image can be overcome to change the shadow brought in imaging process conditional
It rings, the application of correcting algorithm can be overcome, evaded the operation of correction, can also have been generated while having evaded correction calculation
Effect, reliable matching result；
(2) influencing insensitive that is, smaller disturbance to labour smaller in high-definition picture does not influence to screen
It is carried out in journey, this just effectively directly ignores interference data on source, has preferable fault-tolerant ability；
It (3), can be simultaneously to including multiple targets on the basis of retaining legacy data characteristic by the comparison of characteristic point
Image and image in the shielded situation of target part handled, be suitable for the actual conditions in dangerous material stockyard；
(4) not complicated data calculation process, whole process is simple and direct, and it is conventional screening calculation amount that calculating process, which takes,
30%-50%, screening recognition efficiency are well improved.
It is obvious to a person skilled in the art that invention is not limited to the details of the above exemplary embodiments, Er Qie
In the case of without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by appended power
Profit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent requirements of the claims
Variation is included within the present invention.Any reference signs in the claims should not be construed as limiting the involved claims.
Claims (10)
1. a kind of method based on high-definition remote sensing screening dangerous material stockyard, which is characterized in that include the following steps：
S100, target area is obtained based on textural characteristics, distinguishes different textural characteristics by gray level co-occurrence matrixes and structure is special
Property, and target area is extracted according to the uniformity coefficient of gray scale；
S200, feature point extraction is carried out by the method for scale-space, by scale parameter and direction ginseng in target area
Number carries out convolution and responds the invariant features point of extraction image local maximum and minimum point as image as interest；
S300, the space error that invariant features point is eliminated by translation vector carry out model and image by Hausdorff distances
Matching, and by invariant features point carry out translation transformation eliminate space error, matching is fitted to invariant features point；
S400, invariant features point library is established, remote sensing features point is screened one by one, matched inconvenient characteristic point will be fitted successively
Sequence composition invariant features point library, and pass sequentially through contrast characteristic's point and screen, obtain dangerous material stockyard.
2. a kind of method based on high-definition remote sensing screening dangerous material stockyard according to claim 1, which is characterized in that
Gray level co-occurrence matrixes distinguish the specific steps of different textural characteristics and structure feature：
Setting the gray level that L indicates image local area, then gray level co-occurrence matrixes are expressed as the matrix of LxL in the region,
That is Pθ, d=[PI, j]LxL, wherein matrix element PijIt is θ to be expressed as direction in image local area, and distance is d, gray value point
Not Wei i and j pixel to the frequency of appearance；
Co-occurrence matrix feature is extracted by statistical measurement by the frequency of appearance, obtains the grey scale change in the region and texture spy
Sign.
3. a kind of method based on high-definition remote sensing screening dangerous material stockyard according to claim 2, which is characterized in that
Progress textural characteristics are filtered after extracting spatial frequency and location information by gray level co-occurrence matrixes by Gabor finely to extract,
Specific method is：
S101, set the functional form of Gabor filter as Wherein x, y are position coordinates, and j is space domain characteristic value, and w is concussion frequency, σxAnd σyFiltering is indicated respectively
Standard deviation of the device in the directions x and the directions y；
S102, scale parameter u and directioin parameter v is added, then filter function is：
guv(x, y)=a-uG (x ', y '), wherein u, v are integer, a ＞ 1,
And derivative
DerivativeWherein K is total direction number；
S103, by filter function input image I (x, y) carry out m scale, the Gabor transformation in n direction, and to uniformly
The feature of scale different directions is averaged to obtain m textural characteristics subband to get the characteristic vector tieed up to a m as in image
Texture feature information X={ the x of each pixel (x, y)1, x2..., xm}。
4. a kind of method based on high-definition remote sensing screening dangerous material stockyard according to claim 1, which is characterized in that
In step s 200, the specific algorithm of scale-space method is：
S201, the standard deviation Gauss equation for setting corresponding different scale are as L (u, v, σ)=G (u, v, σ) * I (u, v), wherein u
Scale parameter, v are directioin parameter, and σ is scale, and G (u, v, σ) is two-dimensional Gaussian function, and I (u, v) is scale-side of remote sensing images
To function；
S202, two-dimensional Gaussian function is calculated
S203, scale-directivity function I (u, the v)=L (u, v, σ) for setting remote sensing images carry out cycle calculations, until scale σ is equal to
Parameter value, and the reference value of scale σ is 1.6.
5. a kind of method based on high-definition remote sensing screening dangerous material stockyard according to claim 4, which is characterized in that
When carrying out the calculating of scale-space method, two-way calculating should be carried out.
6. a kind of method based on high-definition remote sensing screening dangerous material stockyard according to claim 1, which is characterized in that
The Hausdor distances are specifically defined as：Set two remote sensing features point A={ a1, a2..., apAnd B={ b1, b2...,
bq, then Hausdor distances are H (A, B)=max (h (A, B), h (B, A)), wherein h (A, B)=maxa∈Aminb∈B| | a-b | |, |
| | | it is distance measure.
7. a kind of method based on high-definition remote sensing screening dangerous material stockyard according to claim 6, which is characterized in that
In Hausdor distances calculate, acceleration strategy may be used and calculated in limited range.
8. a kind of method based on high-definition remote sensing screening dangerous material stockyard according to claim 1, which is characterized in that
In step S300, invariant features point carry out translation transformation the specific steps are：
Two S301, setting feature point set A={ a1, a2..., apAnd B={ b1, b2..., bq, and take feature point set be used as away from
Reference set from transformation；
S302, setting D [x, y]=mina∈A| | (x, y)-a | | it is characterized the distance map of point set A, sets D, [x, y]=minb∈B||
(x, y)-b | | it is characterized the distance map of point set B, then there are seek Hausdor distances when translation transformation relative to each other between A and B；
S303, the ratio of defined feature point set A and feature point set B or the ratio of feature point set B and feature point set A are transformation threshold
Value, and it is F (x, y) to set the distances of the opposite Hausdor between A and B, during carrying out translation transformation so that F (x, y)
It is effective translation transformation that the change threshold being minimized, which is less than 1,.
9. a kind of method based on high-definition remote sensing screening dangerous material stockyard according to claim 8, which is characterized in that
After composition invariant features point library, specific screening step is as follows：
S401, setting time coefficient are γ1, cloud amount coefficient is γ2, resolution rate coefficient is γ3, wherein 0≤γ≤1, and γ1+γ2
+γ3=1；
S402, remote sensing image weighted value be γ=γ1xλ1+γ2xλ2+γ3xλ3, wherein λ1、λ2And λ3Respectively coefficient of correspondence
Weight；
S403, the weighted value that the record per data is calculated according to the weighted value of above-mentioned remote sensing image, and to all remotely-sensed datas
It is screened, the maximum data of weighted value is charged into filter record collection.
10. a kind of method based on high-definition remote sensing screening dangerous material stockyard according to claim 1, feature exist
In carrying out duplicate removal processing to the filter record collection of acquisition, obtain new result set, retrieve and judge the data that filter record is concentrated
Whether in result set, there is no result set is then added, until completing the retrieval of all data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711312959.9A CN108304766B (en) | 2017-12-12 | 2017-12-12 | A method of dangerous material stockyard is screened based on high-definition remote sensing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711312959.9A CN108304766B (en) | 2017-12-12 | 2017-12-12 | A method of dangerous material stockyard is screened based on high-definition remote sensing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108304766A true CN108304766A (en) | 2018-07-20 |
| CN108304766B CN108304766B (en) | 2019-01-04 |
Family
ID=62869932
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711312959.9A Active CN108304766B (en) | 2017-12-12 | 2017-12-12 | A method of dangerous material stockyard is screened based on high-definition remote sensing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108304766B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109977909A (en) * | 2019-04-04 | 2019-07-05 | 山东财经大学 | Finger vein identification method and system based on minutiae point Region Matching |
| CN110516531A (en) * | 2019-07-11 | 2019-11-29 | 广东工业大学 | A kind of recognition methods of the dangerous mark based on template matching |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100182480A1 (en) * | 2009-01-16 | 2010-07-22 | Casio Computer Co., Ltd. | Image processing apparatus, image matching method, and computer-readable recording medium |
| CN103020945A (en) * | 2011-09-21 | 2013-04-03 | 中国科学院电子学研究所 | Remote sensing image registration method of multi-source sensor |
| CN103793705A (en) * | 2014-03-11 | 2014-05-14 | 哈尔滨工业大学 | Non-contact palm print authentication method based on iterative random sampling consistency algorithm and local palm print descriptor |
| CN103839265A (en) * | 2014-02-26 | 2014-06-04 | 西安电子科技大学 | SAR image registration method based on SIFT and normalized mutual information |
| US20140270411A1 (en) * | 2013-03-15 | 2014-09-18 | Henry Shu | Verification of User Photo IDs |
| CN104318548A (en) * | 2014-10-10 | 2015-01-28 | 西安电子科技大学 | Rapid image registration implementation method based on space sparsity and SIFT feature extraction |
| CN104408701A (en) * | 2014-12-03 | 2015-03-11 | 中国矿业大学 | Large-scale scene video image stitching method |
| CN104504724A (en) * | 2015-01-15 | 2015-04-08 | 杭州国策商图科技有限公司 | Moving object extracting and tracking algorithm capable of being not affected by obstacles |
-
2017
- 2017-12-12 CN CN201711312959.9A patent/CN108304766B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100182480A1 (en) * | 2009-01-16 | 2010-07-22 | Casio Computer Co., Ltd. | Image processing apparatus, image matching method, and computer-readable recording medium |
| CN103020945A (en) * | 2011-09-21 | 2013-04-03 | 中国科学院电子学研究所 | Remote sensing image registration method of multi-source sensor |
| US20140270411A1 (en) * | 2013-03-15 | 2014-09-18 | Henry Shu | Verification of User Photo IDs |
| CN103839265A (en) * | 2014-02-26 | 2014-06-04 | 西安电子科技大学 | SAR image registration method based on SIFT and normalized mutual information |
| CN103793705A (en) * | 2014-03-11 | 2014-05-14 | 哈尔滨工业大学 | Non-contact palm print authentication method based on iterative random sampling consistency algorithm and local palm print descriptor |
| CN104318548A (en) * | 2014-10-10 | 2015-01-28 | 西安电子科技大学 | Rapid image registration implementation method based on space sparsity and SIFT feature extraction |
| CN104408701A (en) * | 2014-12-03 | 2015-03-11 | 中国矿业大学 | Large-scale scene video image stitching method |
| CN104504724A (en) * | 2015-01-15 | 2015-04-08 | 杭州国策商图科技有限公司 | Moving object extracting and tracking algorithm capable of being not affected by obstacles |
Non-Patent Citations (2)
| Title |
|---|
| 孔韦韦: "《图像融合技术 基于多分辨率非下采样理论与方法》", 31 July 2015, 西安：西安电子科技大学出版社 * |
| 赵荣椿等: "《精通图像处理经典算法 MATLAB版》", 30 April 2014, 北京：北京航空航天大学出版社 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109977909A (en) * | 2019-04-04 | 2019-07-05 | 山东财经大学 | Finger vein identification method and system based on minutiae point Region Matching |
| CN109977909B (en) * | 2019-04-04 | 2021-04-20 | 山东财经大学 | Finger vein identification method and system based on minutia area matching |
| CN110516531A (en) * | 2019-07-11 | 2019-11-29 | 广东工业大学 | A kind of recognition methods of the dangerous mark based on template matching |
| CN110516531B (en) * | 2019-07-11 | 2023-04-11 | 广东工业大学 | Identification method of dangerous goods mark based on template matching |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108304766B (en) | 2019-01-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110287932B (en) | Road blocking information extraction method based on deep learning image semantic segmentation | |
| WO2018024030A1 (en) | Saliency-based method for extracting road target from night vision infrared image | |
| Xu et al. | Object‐based mapping of karst rocky desertification using a support vector machine | |
| CN107067405B (en) | Remote sensing image segmentation method based on scale optimization | |
| CN110765934B (en) | Geological disaster identification method based on multi-source data fusion | |
| CN111476170A (en) | Remote sensing image semantic segmentation method combining deep learning and random forest | |
| CN108765359A (en) | A kind of fusion method of target in hyperspectral remotely sensed image and full-colour image based on JSKF models and NSCT technologies | |
| CN109063754A (en) | A kind of remote sensing image multiple features combining classification method based on OpenStreetMap | |
| Guan et al. | Random forests-based feature selection for land-use classification using lidar data and orthoimagery | |
| CN110363236B (en) | Hyperspectral image extreme learning machine clustering method for embedding space-spectrum combined hypergraph | |
| WO2022141145A1 (en) | Object-oriented high-resolution remote sensing image multi-scale segmentation method and system | |
| Gueguen et al. | Toward a generalizable image representation for large-scale change detection: Application to generic damage analysis | |
| CN108304766B (en) | A method of dangerous material stockyard is screened based on high-definition remote sensing | |
| CN110363053A (en) | A kind of Settlement Place in Remote Sensing Image extracting method and device | |
| Baraldi et al. | Operational performance of an automatic preliminary spectral rule-based decision-tree classifier of spaceborne very high resolution optical images | |
| Li et al. | A pseudo-siamese deep convolutional neural network for spatiotemporal satellite image fusion | |
| CN110309763A (en) | A kind of building classification method based on satellite remote sensing images | |
| Zhang et al. | Watershed segmentation algorithm based on LUV color space region merging for extracting slope hazard boundaries | |
| HUAN XIE et al. | A new binary encoding algorithm for the simultaneous region-based classification of hyperspectral data and digital surface models | |
| CN111460943A (en) | Remote sensing image ground object classification method and system | |
| CN116091930A (en) | Terrace drawing method, device, equipment and storage medium based on remote sensing | |
| CN113378924B (en) | Remote sensing image supervision and classification method based on space-spectrum feature combination | |
| CN114862883A (en) | Target edge extraction method, image segmentation method and system | |
| CN115147726A (en) | City form map generation method and device, electronic equipment and readable storage medium | |
| CN108764145A (en) | One kind is towards Dragon Wet Soil remote sensing images density peaks clustering method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |