CN109584262A - Cloud detection method of optic, device and electronic equipment based on remote sensing image - Google Patents
Cloud detection method of optic, device and electronic equipment based on remote sensing image Download PDFInfo
- Publication number
- CN109584262A CN109584262A CN201811536035.1A CN201811536035A CN109584262A CN 109584262 A CN109584262 A CN 109584262A CN 201811536035 A CN201811536035 A CN 201811536035A CN 109584262 A CN109584262 A CN 109584262A
- Authority
- CN
- China
- Prior art keywords
- remote sensing
- sensing image
- connected region
- preliminary
- bipartite graph
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The present invention is applicable in cloud detection technical field, provides a kind of cloud detection method of optic based on remote sensing image, device and electronic equipment, this method comprises: carrying out binaryzation label to remote sensing image using threshold method, obtains preliminary bipartite graph；Connected region is carried out to the preliminary bipartite graph to identify to obtain three component of connected region；Cloud detection is carried out to three component of connected region, obtains the alpha figure of the remote sensing image, the case where realizing and efficiently carry out cloud detection to remote sensing image, and improve the accuracy of cloud detection, avoid part missing inspection, false retrieval.
Description
Technical field
The invention belongs to cloud detection technical fields, more particularly to the cloud detection method of optic based on remote sensing image, device and electronics
Equipment.
Background technique
As the technology of remote sensing satellite constantly develops, remote sensing technology plays in meteorological scientific research with application aspect more next
More important role.Its medium cloud is the image factor important in meteorological and climatic study, is obtained in wind and cloud remote sensing image
Cheng Zhong is influenced by factors such as cloud and mist interference, leads to original place object light spectrum distortion, Remote Sensing Products and image interpretation are influenced, to information
Extraction causes very big influence.Being correctly separated in remote sensing image has cloud pixel and cloudless pixel, for weather forecast, meteorology
Prevention, temperature retrieval, rescue and monitoring of ecological environment of disaster etc. have great influence, therefore satellite cloud picture cloud detection becomes
It obtains more and more important.Carrying out cloud detection work can be the application of the rejecting of subsequent cloud, cloud classification and other field remotely-sensed data
It lays a solid foundation.
Traditional cloud detection method of optic mainly based on threshold segmentation method, what cloud showed in visible light and near infrared spectrum
Characteristic is the low bright temperature of high reflectance, using this characteristic, can choose suitable threshold value and determines in pixel whether there is cloud.Current
Threshold method includes: spectrum combination threshold method, frequency combination threshold method.
Spectrum combination threshold method mainly utilizes cloud to have the characteristic of strong reflection in visible light wave range, such algorithm is to threshold value
Sensitivity is higher, and same satellite data increases such because great variety will occur for the reasons such as time, weather, detection threshold value
The limitation of method.Frequency combination threshold method mainly utilizes the low frequency characteristic of cloud, passes through the methods of wavelet analysis, Fourier transformation
It obtains image low-frequency data and carries out cloud detection, but due to generalling use multi-level Wavelet Transform transformation and excluding by the interference of ground low-frequency information,
This greatly reduces cloud detection efficiency.
Another kind of tradition cloud detection method of optic is mainly classified according to physical features such as the texture of cloud, shape, gray scales.
Texture analysis method is using cloud and ground texture feature difference, often as unit of piecemeal subgraph, in conjunction with second moment, fractal dimension, ash
It spends co-occurrence matrix and multiple bilateral filtering carries out textural characteristics calculating, such method needs to obtain reliable cloud characteristic interval in advance
It can guarantee that the precision of classification, efficiency are lower.Statistical method is broadly divided into statistic equation and clustering methodology.Statistic equation method benefit
The reflectivity for simulating formula calculating cloud or bright temperature are established with sample data to carry out cloud detection, clustering methodology is according to difference
There are the principles of apparent difference to realize that cloud detection will be clustered when sample size is larger for the pixel observation of type of ground objects
Conclusion has certain difficulty, needs human intervention, extreme influence detection efficiency.
Comprehensive intelligent method also gradually increases in the application of remote sensing fields in recent years, and comprehensive a variety of comprehensive intelligent methods mainly include
Artificial neural network, support vector machines and fuzzy logic algorithm etc..Comprehensive intelligent method needs to obtain during realization largely
Training sample, it is more demanding to the selection of characteristic of division, need to choose sample again for different data, lead to low efficiency
Under.
Summary of the invention
The purpose of the present invention is to provide cloud detection method of optic, device and electronic equipments based on remote sensing image, it is intended to solve
Since the prior art can not efficiently and accurately carry out the technical issues of cloud detection.
In a first aspect, the method includes following steps the present invention provides a kind of cloud detection method of optic based on remote sensing image
It is rapid:
Binaryzation label is carried out to remote sensing image using threshold method, obtains preliminary bipartite graph；
Connected region is carried out to the preliminary bipartite graph to identify to obtain three component of connected region；
Cloud detection is carried out to three component of connected region, obtains the alpha figure of the remote sensing image.
Optionally, the step of being carried out by binaryzation label, obtains preliminary bipartite graph for remote sensing image using threshold method, comprising:
Calculate separately gray threshold of the remote sensing image under a variety of threshold methods；
For each pixel of remote sensing image, binaryzation mark is carried out to the pixel according to the gray threshold under each threshold method
Note；
It marks to obtain preliminary bipartite graph by the binaryzation of each pixel.
Optionally, connected region is carried out to the preliminary bipartite graph and identifies the step of obtaining three component of connected region, comprising:
The preliminary bipartite graph is progressively scanned using Run- Length Coding, record in every a line starting point that each rolls into a ball,
Terminal and line number；
The continuous group of scanning, it is set in equivalent table initial markers and record it is of equal value right；
Traversal starts the label of group, and the equivalence searched in the equivalent table is right, assigns of equal value to same label；
The label of each group is inserted in the preliminary bipartite graph, connected component labeling figure is obtained, with the weight of connected region
Heart point is seed point, is set as determining prospect, obtains three component of connected region.
Optionally, cloud detection is carried out to three component of connected region, obtains the step of the alpha figure of the remote sensing image
Suddenly, comprising:
According to the relationship of each pixel and its neighborhood, increases mobile weight, determine that two multiply Laplacian Matrix；
According to the weight, nomography is scratched using the form of closing, cloud detection is carried out to three component of connected region, obtain institute
State the alpha figure of remote sensing image.
Optionally, according to the weight, nomography is scratched using the form of closing, cloud detection is carried out to three component of connected region,
The step of obtaining the alpha figure of the remote sensing image, comprising:
The least square Laplacian Matrix in nomography is scratched according to the weight and form of closing, constructs the remote sensing image
Minimum movement two multiply Laplacian Matrix；
According to the mobile Laplacian Matrix, schemed using the alpha that conjugate gradient algorithms solve the remote sensing image.
Second aspect provides a kind of cloud detection device based on remote sensing image, comprising:
Binaryzation mark module obtains preliminary bipartite graph for carrying out binaryzation label to remote sensing image using threshold method；
Connected region identification module identifies to obtain connected region three for carrying out connected region to the preliminary bipartite graph and divide
Figure；
Cloud detection module obtains the alpha of the remote sensing image for carrying out cloud detection to three component of connected region
Figure.
Optionally, the binaryzation mark module includes:
Gray threshold computing unit, for calculating separately gray threshold of the remote sensing image under a variety of threshold methods；
Binaryzation marking unit, for being directed to each pixel of remote sensing image, according to the gray threshold pair under each threshold method
The pixel carries out binaryzation label；
Bipartite graph generation unit obtains preliminary bipartite graph for marking by the binaryzation of each pixel.
Optionally, the connected region identification module includes:
Group's marking unit records each for carrying out progressive scan group to the preliminary bipartite graph using Run- Length Coding
The starting point of group, terminal and line number；
Equivalence is to recording unit, for scanning continuous group, sets initial markers to it in equivalent table and records equivalence
It is right；
Group's marking unit searches the equivalent sequence in equivalent table, it is same to assign equivalent sequence for traversing the label for starting group
Sample label；
Connected component labeling unit obtains connected region for inserting the label of each group in the preliminary bipartite graph
Label figure is set as determining prospect, obtains three component of connected region using the focus point of connected region as seed point.
The third aspect provides a kind of electronic equipment, comprising:
Processor；And
The memory being connect with the processor communication；Wherein,
The memory is stored with readable instruction, and the readable instruction realizes such as first when being executed by the processor
Method described in aspect.
Fourth aspect provides a kind of computer readable storage medium, is stored thereon with computer program, the calculating
Machine program realizes the method such as first aspect when executed.
The present invention carries out binaryzation label certainly to remote sensing image when carrying out cloud detection based on remote sensing image, using threshold method
It is dynamic to generate preliminary bipartite graph, then connected region is carried out to the preliminary bipartite graph and identifies to obtain three component of connected region, to connection
Three component of region carries out cloud detection and obtains the alpha figure of remote sensing image, realizes and efficiently carries out cloud detection to remote sensing image, and mentions
The high accuracy of cloud detection, the case where avoiding part missing inspection, false retrieval.
Detailed description of the invention
Fig. 1 is the implementation flow chart for the cloud detection method of optic based on remote sensing image that the embodiment of the present invention one provides；
Fig. 2 is the schematic diagram of the preliminary bipartite graph (a) and three component (b) of connected region shown in the embodiment of the present invention one；
Fig. 3 is the structural block diagram of the cloud detection device provided by Embodiment 2 of the present invention based on remote sensing image；
Fig. 4 is the structural block diagram for the electronic equipment 100 that the embodiment of the present invention three provides；
Fig. 5 is the offer of the embodiment of the present invention four to a kind of FY-2G image progress schematic diagram of cloud detection；
Fig. 6 is the offer of the embodiment of the present invention four to FY-2G image progress cloud detection another kind schematic diagram；
Fig. 7 is the offer of the embodiment of the present invention four to a kind of Landsat8 image progress schematic diagram of cloud detection；
Fig. 8 is the offer of the embodiment of the present invention four to Landsat8 image progress cloud detection another kind schematic diagram.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
Specific implementation of the invention is described in detail below in conjunction with specific embodiment:
Embodiment one:
Fig. 1 shows the implementation process of the cloud detection method of optic based on remote sensing image of the offer of the embodiment of the present invention one.This hair
Bright embodiment is suitable for the electronic equipments such as smart phone, computer, processor is arranged in these electronic equipments, according to the distant of input
Feel data and carries out cloud detection.For ease of description, only parts related to embodiments of the present invention are shown, and details are as follows:
In step S110, binaryzation label is carried out to remote sensing image using threshold method, obtains preliminary bipartite graph.
Remote sensing image (RS, Remote Sensing Image) refers to the film or photograph for recording various atural object electromagnetic wave sizes
Piece is broadly divided into airphoto and satellite photograph.Remote sensing image in the present invention is to image captured by cloud.
Normally, manual markings need to be carried out to the remote sensing image of input by scratching drawing method due to the form of closing, and be to avoid craft
The fault and workload of label carry out binaryzation label to remote sensing image by advancing with threshold method in the present embodiment, automatically
Generate preliminary bipartite graph.
The accuracy that binaryzation label is carried out to further increase threshold method to remote sensing image, avoids because of single threshold value method
Defect causes binaryzation label deviation occur, passes through in the present embodiment and gathers multiple threshold methods to remote sensing image progress binaryzation mark
Note.
Optionally, multiple threshold methods can carry out comprehensive descision by a kind of combination strategy.For example, basic threshold method packet
Include: Da-Jin algorithm, piecemeal Da-Jin algorithm, threshold method, global threshold and local threshold combined techniques etc., Wellner are adaptive
Threshold value, minimum error method, Two-peak method, iteration method, maximum entropy threshold method, fixed threshold split plot design etc..In order to as far as possible
Keep preliminary bipartite graph of the invention more accurate, the mode that the present invention has selected temporal voting strategy to integrate as threshold value result.That is, right
In a width original remote sensing image, a variety of threshold values are generated using multi-threshold method as a result, a ballot value is set, when each threshold value result
Votes are more than or equal to the votes, then the original state of the pixel is set as 1, is otherwise set as 0.And then pass through each pixel
Binaryzation marks to obtain preliminary bipartite graph.
It is understood that binaryzation label can also be carried out using other temporal voting strategies and obtain preliminary bipartite graph, example
Such as, ballot ratio can be set, when each threshold value result ballot ratio is more than or equal to the ballot ratio, then the pixel is initial
State is set as 1, is otherwise set as 0.
In step S120, connected region is carried out to preliminary bipartite graph and identifies to obtain three components.
The label rolled into a ball using Run- Length Coding to preliminary bipartite graph, progressive scan group, records starting point, the terminal of each group
And line number；The continuous group of scanning, it is set in equivalent table initial markers and record it is of equal value right；Analytic equivalence class, search etc.
Equivalent sequence in valence class assigns equivalent sequence to same label；Finally by the label filling tag image of each group, obtain
To connected region image, then using the focus point of connected region as seed point, and it is set to prospect, obtains corresponding three points
Figure.
It for preliminary bipartite graph, is scanned line by line, the sequence of continuous gray pixels composition is known as group in a line, and remembers
The starting point of every a line, terminal, line number under record.After scanned, all groups of full figure are obtained, and recorded specific position
It sets.During subsequent connected component labeling, different groups is classified as equivalence class and parity price class makes marks and requires to know
The specific location of group.
Run- Length Coding by image tagged at a series of group, the group in all rows other than the first row, if it is with before
All groups in a line then give its new label all without overlapping region；If it only has weight with a group in lastrow
Region is closed, then the label of that group of lastrow is assigned to it；If there are overlapping region in 2 or more groups of it and lastrow,
The minimum label of the group that is connected then is assigned to current group, and the label write-in of these groups of lastrow is of equal value right, illustrate it
Belong to one kind.By equivalence to the equivalent sequence be converted in equivalent table, each sequence is needed to an identical label (from 1
Start, give one label of each equivalent sequence).
In equivalence class, the label of each equivalent sequence be it is identical, for different groups, be clear which roll into a ball
Belong to one kind, facilitates and equivalent sequence is done into same class label.
The label of the group of traversal searches the equivalent sequence in equivalence class, assigns equivalent sequence to same label.It finally will be every
In the label filling tag image of a group, connected region figure is obtained, then using connected region focus point as seed point, be set as determining
Prospect generates three last components.
In order to guarantee image detail, the present invention uses 4 neighborhood connected component labelings.As shown in table 1,4 neighborhoods are to center
0,1,2,3 point C neighbouring positions are judged, if there are the pixel labeled as 1 in 0,1,2,3 positions, then it is assumed that C point and 0,1,2,
3 location points are connected.Otherwise it is assumed that C point is isolated point.Table 2 (a) is marked using 4 neighborhood connected component labelings, obtains table
2 (b) share 4 regions.The label that each connected region marks each connected region later is obtained, corresponding connected region is obtained
Three components, such as table 2 (b).
Table 1
Table 2 (a)
Table 2 (b)
Fig. 2 is the schematic diagram of the preliminary bipartite graph (a) and three components (b) shown in the present embodiment.
In step S130, cloud detection is carried out to three component of connected region, obtains the alpha figure of remote sensing image.
According to the relationship of each pixel and its neighborhood, increases mobile weight, determine that two multiply Laplacian Matrix, moved according to minimum
Dynamic Laplacian Matrix scratches nomography using the form of closing and carries out cloud detection to three component of connected region, obtains remote sensing image
Alpha figure.
The original stingy drawing method of form that closes needs to solve Large Scale Sparse Linear equation, therefore inefficient.Close formal approach
Based on local linear it is assumed that being expressed as follows:
αi=aIi+b,i∈wi (1)
When assumed condition is invalid in local neighborhood, especially neighborhood is bigger and effect is not in the case that texture is complicated
It is good.Assuming that alpha value meets linear conditions in neighborhood, least square method solution part is used different from original formal approach of closing
Linear relationship, the present invention is in window wiIt is interior to solve local linear relationship using Moving Least, it is expressed as follows:
Figure is scratched with the form of closing to be different in: weight ω, mobile minimum two are increased in minimizing formula (2) and formula (3)
Multiply smaller in the local weight ω remoter apart from current pixel, therefore Moving Least can solve more accurate local line
Sexual intercourse, it is more more effective than the linear relationship that least square method solves.ωiIt is neighborhood wkWeight.Formula (2) and formula (3) can be with tables
It is shown as the form of following matrix:
For each neighborhood wk, GkIt is defined as | | wk| | × 2 matrixes.GkEvery row includes vector (Ii,1)。WkIt is every row vector
Corresponding weight vector.Gk' it is GkWkWeighting, corresponding every row vector indicate (Wk·Ii,Wk),It is all pixels in neighborhood
The vector of corresponding alpha value composition.
Coefficient ak,bkIt solves as follows
It enablesJ (α) is represented by
δ in formulai,jIt is Kronecker delta function；μkAnd σ2It is wicket w respectivelykIt is interior based on WkWeighted mean
And variance；||wk| | it is the number of pixel in window.
It is similar under color model and closes form algorithm, the linear relationship of each interchannel of color image is indicated with formula (9)
After carrying out abbreviation to formula (10), it is as follows to solve mobile Laplacian Matrix under color model:
J (α)=α L αT (11)
Since in improved stingy figure cloud detection algorithm, if the size of core is r, image pixel number is imagesize, storage
Space complexity required for Laplacian Matrix L is imagesize*r2, calculating space complexity urgency with the increase of core
Increase severely big.The art of computation in big kernel method is used for reference, improved conjugate gradient method solution value is used.For equation Lx=b, conjugation
The key of gradient method is to construct conjugate vector p, and seeks its corresponding residual error.Conjugate gradient method can be solved with alternative manner.
In each iterative process, new conjugate vector is as follows by solving:
The coefficient of conjugate direction solves as follows:
New x value and residual error is as follows with solving:
xk=xk-1+skpk (15)
rk=rk-1+skLpk (16)
Crucial step is to solve vector Lp in conjugate gradient solution procedure, and the spatial complex of direct solution L is spent
Greatly, but the dimension of Lp is imagesize, it is therefore desirable to avoid direct solution L, and directly solve i pairs of Lp vector midpoint with following formula
The element q answeredi。
In formula: WkIt is the corresponding neighborhood of pixel k；||wk| | it is the size of neighborhood；I is to surround pixel k neighborhood WkIn one
Pixel；qiFor i-th of element of q vector；IiFor corresponding 3 dimensional vector of pixel i, R, tri- channels G, B are indicated；piFor conjugation to
The corresponding element of pixel i in amount；μkIt is 3 dimensional vectors, is neighborhood WkMiddle IiThe mean value of vector；For neighborhood WkMiddle element i is corresponding
Conjugate vector piMean value；It is corresponding 3 dimensional vector of pixel k；For the corresponding scalar of pixel k.
Solution obtains the corresponding element q of Lp vector midpoint ii, Lp vector is constituted, formula (16) is brought into and obtains new residual error
Value, iteration acquire the alpha value that x value is this paper.Alpha value is cloud detection result.
(Lp)iThe correctness of calculation formula is guaranteed by following theorem:
Theorem 1: (Lp) that formula (17) is calculatediIt is calculated (Lp) with utilization formula (12)iIt is of equal value.
It proves: enabling q=Lp, since q and p is linear relationship, only need to prove formula (21)
Formula (20) are substituted into formula (17) to eliminateIt is available:
In addition, having
According to formula (18), and to pjLocal derviation is done, can be obtained
Formula (23) and formula (24) are substituted into formula (22), can be obtained
Corresponding Laplacian Matrix L in formula (25) i.e. formula (12).
Using method as described above, when carrying out cloud detection based on remote sensing image, using threshold method to remote sensing image into
Row binaryzation label automatically generates preliminary bipartite graph, then carries out connected region to the preliminary bipartite graph and identify to obtain connected region
Three components, to three component of connected region carry out cloud detection obtain remote sensing image alpha scheme, realize efficiently to remote sensing image into
The case where racking detection, and improving the accuracy of cloud detection, avoid part missing inspection, false retrieval.
Embodiment two:
Fig. 3 shows the structural block diagram of the cloud detection device provided by Embodiment 2 of the present invention based on remote sensing image, in order to
Convenient for explanation, only parts related to embodiments of the present invention are shown, including:
Binaryzation mark module 110 obtains preliminary two points for carrying out binaryzation label to remote sensing image using threshold method
Figure；
Connected region identification module 120 identifies to obtain connected region three for carrying out connected region to preliminary bipartite graph and divide
Figure；
Cloud detection module 130 obtains the alpha figure of remote sensing image for carrying out cloud detection to three component of connected region.
Preferably, binaryzation mark module 110 includes:
Gray threshold computing unit 111, for calculating separately gray threshold of the remote sensing image under a variety of threshold methods；
Binaryzation marking unit 112, for being directed to each pixel of remote sensing image, according to the gray threshold under each threshold method
Binaryzation label is carried out to pixel；
Bipartite graph generation unit 113 obtains preliminary bipartite graph for marking by the binaryzation of each pixel.
Preferably, connected region identification module 120 includes:
Group's marking unit 121 records each for carrying out progressive scan group to the preliminary bipartite graph using Run- Length Coding
The starting point of a group, terminal and line number；
Equivalence is to recording unit 122, for scanning continuous group, set initial markers to it in equivalent table and record etc.
Valence pair；
Marking unit 123 is traversed, for traversing the label for starting group, the equivalent sequence in equivalent table is searched, assigns of equal value
Sequence equally marks；
Connected component labeling unit 124 obtains connected region for inserting the label of each group in the preliminary bipartite graph
Field mark figure is set as determining prospect, obtains three components using the focus point of connected region as seed point.
Preferably, cloud detection module 130 includes:
Weight determination unit 131, at a distance from current pixel, determining that two multiply Laplacian Matrix according to each pixel
Weight；
Cloud detection unit 132, for scratching nomography using the form of closing and carrying out cloud inspection to three component of connected region according to weight
It surveys, obtains the alpha figure of remote sensing image.
In embodiments of the present invention, should each module of cloud detection device based on remote sensing image can be by corresponding hardware or soft
Part unit realizes that each module can be independent soft and hardware module, also can integrate as a soft and hardware unit, does not have to herein
To limit the present invention.The specific embodiment of each module can refer to the description of embodiment one, and details are not described herein.
Embodiment three:
The structural block diagram that Fig. 4 shows the electronic equipment 100 of the offer of the embodiment of the present invention three is only shown for ease of description
Part related to the embodiment of the present invention is gone out.
With reference to Fig. 4, electronic equipment 100 may include one or more following component: processing component 101, memory
102, power supply module 103, multimedia component 104, audio component 105, sensor module 107 and communication component 108.Wherein,
Said modules and be not all it is necessary, electronic equipment 100 can according to itself functional requirement increase other assemblies or reduce it is certain
Component, this embodiment is not limited.
The integrated operation of the usual controlling electronic devices 100 of processing component 101, it is such as logical with display, call, data
Letter, camera operation and the associated operation of record operation etc..Processing component 101 may include one or more processors 109
It executes instruction, to complete all or part of the steps of aforesaid operations.In addition, processing component 101 may include one or more
Module, convenient for the interaction between processing component 101 and other assemblies.For example, processing component 101 may include multi-media module,
To facilitate the interaction between multimedia component 104 and processing component 101.
Memory 102 is configured as storing various types of data to support the operation in electronic equipment 100.These data
Example include any application or method for being operated on electronic equipment 100 instruction.Memory 102 can be by appointing
The volatibility or non-volatile memory device or their combination of what type are realized, such as SRAM (Static Random
Access Memory, static random access memory), EEPROM (Electrically Erasable Programmable
Read-Only Memory, electrically erasable programmable read-only memory), EPROM (Erasable Programmable Read
Only Memory, Erasable Programmable Read Only Memory EPROM), (Programmable Read-Only Memory may be programmed PROM
Read-only memory), ROM (Read-Only Memory, read-only memory), magnetic memory, flash memory, disk or CD.
One or more modules are also stored in memory 102, which is configured to be handled by the one or more
Device 109 executes, to complete all or part of step in following any shown method.
Power supply module 103 provides electric power for the various assemblies of electronic equipment 100.Power supply module 103 may include power supply pipe
Reason system, one or more power supplys and other with for electronic equipment 100 generate, manage, and distribute the associated component of electric power.
Multimedia component 104 includes the screen of one output interface of offer between electronic equipment 100 and user.One
In a little embodiments, screen may include LCD (Liquid Crystal Display, liquid crystal display) and TP (Touch
Panel, touch panel).If screen includes touch panel, screen may be implemented as touch screen, from the user to receive
Input signal.Touch panel includes one or more touch sensors to sense the gesture on touch, slide, and touch panel.Institute
The boundary of a touch or slide action can not only be sensed by stating touch sensor, but also be detected and the touch or slide phase
The duration and pressure of pass.
Audio component 105 is configured as output and/or input audio signal.For example, audio component 105 includes a Mike
Wind, when electronic equipment 100 is in operation mode, when such as call mode, recording mode, and voice recognition mode, microphone is configured
To receive external audio signal.The received audio signal can be further stored in memory 102 or via communication component
108 send.In some embodiments, audio component 105 further includes a loudspeaker, is used for output audio signal.
Sensor module 107 includes one or more sensors, for providing the state of various aspects for electronic equipment 100
Assessment.For example, sensor module 107 can detecte the state that opens/closes of electronic equipment 100, the relative positioning of component is passed
The coordinate that sensor component 107 can also detect 100 1 components of electronic equipment 100 or electronic equipment changes and electronic equipment
100 temperature change.In some embodiments, which can also include Magnetic Sensor, pressure sensor or temperature
Spend sensor.
Communication component 108 is configured to facilitate the communication of wired or wireless way between electronic equipment 100 and other equipment.
Electronic equipment 100 can access the wireless network based on communication standard, such as WiFi (Wireless-Fidelity, wireless network),
2G or 3G or their combination.In one exemplary embodiment, communication component 108 is received via broadcast channel from outside
The broadcast singal or broadcast related information of broadcasting management systems.In one exemplary embodiment, the communication component 108 also wraps
NFC (Near Field Communication, near-field communication) module is included, to promote short range communication.For example, NFC module can
Based on RFID (Radio Frequency Identification, radio frequency identification) technology, IrDA (Infrared Data
Association, Infrared Data Association) technology, UWB (Ultra-Wideband, ultra wide band) technology, BT (Bluetooth, it is blue
Tooth) technology and other technologies realize.
In the exemplary embodiment, electronic equipment 100 can be by one or more ASIC (Application Specific
Integrated Circuit, application specific integrated circuit), DSP (Digital Signal Processing, at digital signal
Manage device), PLD (Programmable Logic Device, programmable logic device), FPGA (Field-Programmable
Gate Array, field programmable gate array), controller, microcontroller, microprocessor or other electronic components realize, be used for
Execute the above method.
The concrete mode that processor executes operation in server in the embodiment is somebody's turn to do related based on remote sensing image
It is described in detail in the embodiment of cloud detection method of optic, will no longer elaborate explanation herein.The foregoing is merely the present invention
Preferred embodiment, be not intended to limit the invention, it is done within the spirit and principles of the present invention it is any modification,
Equivalent replacement and improvement etc., should all be included in the protection scope of the present invention.
Example IV:
In the present embodiment, effectiveness of the invention will be analyzed by actual experiment data.
It include the validity of image test the method for the present invention of cloud using several.By the cloud detection result of wind and cloud image and greatly
Saliva method (OTSU), national weather satellite center (NSMC) cloud detection compare.Landsat8 image cloud detection result with
OTSU and manual markings cloud atlas are compared.
Fig. 5, Fig. 6 are the cloud detection of FY-2G image, in Fig. 5 and Fig. 6, (a), (b), (c) (d) be respectively visible images,
The testing result schematic diagram that OTSU cloud detection, NSMC and the present invention detect.As can be seen that OTSU cloud detection is lost part carefully
Section, scattered zonule cloud are not detected；NSMC cloud detection main problem is examine more.
Fig. 7, Fig. 8 are the cloud detection of Landsat8 image as a result, in Fig. 7 and Fig. 8, and (a), (b), (c) (d) are respectively original
Figure obtains the testing result schematic diagram that preliminary bipartite graph, OTSU and the present invention detect using manual markings.As seen from the figure, OTSU
There is the case where more inspections in cloud detection result, and method of the present invention is capable of detecting when scattered cloud, retains relatively complete in terms of details
It is whole.I.e. cloud can be effectively detected in the method for the present invention, improve cloud detection efficiency and accuracy rate.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Made any modifications, equivalent replacements, and improvements etc., should all be included in the protection scope of the present invention within mind and principle.
Claims (10)
1. a kind of cloud detection method of optic based on remote sensing image, which is characterized in that the method includes the following steps:
Binaryzation label is carried out to remote sensing image using threshold method, obtains preliminary bipartite graph；
Connected region is carried out to the preliminary bipartite graph to identify to obtain three component of connected region；
Cloud detection is carried out to three component of connected region, obtains the alpha figure of the remote sensing image.
2. the method as described in claim 1, which is characterized in that carry out binaryzation label to remote sensing image using threshold method, obtain
The step of to preliminary bipartite graph, comprising:
Calculate separately gray threshold of the remote sensing image under a variety of threshold methods；
For each pixel of remote sensing image, binaryzation label is carried out to the pixel according to the gray threshold under each threshold method；
By being voted to obtain preliminary bipartite graph to each pixel binary map.
3. the method as described in claim 1, which is characterized in that identify and connected to the preliminary bipartite graph progress connected region
The step of logical three component of region, comprising:
The preliminary bipartite graph is progressively scanned using Run- Length Coding, records starting point, the terminal that each in every a line is rolled into a ball
And line number；
The continuous group of scanning, it is set in equivalent table initial markers and record it is of equal value right；
Traversal starts the label of group, and the equivalence searched in the equivalent table is right, assigns of equal value to same label；By the mark of each group
Note is inserted in the preliminary bipartite graph, is obtained connected component labeling figure, using the focus point of connected region as seed point, is set as true
Determine prospect, obtains three components.
4. the method as described in claim 1, which is characterized in that carry out cloud detection to three component of connected region, obtain institute
The step of stating the alpha figure of remote sensing image, comprising:
According to the relationship of each pixel and its neighborhood, increase mobile weight, building two multiplies Laplacian Matrix；
According to the weight, nomography is scratched using the form of closing, cloud detection is carried out to three component of connected region, obtained described distant
Feel the alpha figure of image.
5. method as claimed in claim 4, which is characterized in that according to the weight, scratch nomography to described using the form of closing
The step of three component of connected region carries out cloud detection, obtains the alpha figure of the remote sensing image, comprising:
The least square Laplacian Matrix in nomography is scratched according to the weight and form of closing, constructs the remote sensing image most
Small movement two multiplies Laplacian Matrix；
Multiply Laplacian Matrix according to the minimum movement two, the alpha of the remote sensing image is solved using conjugate gradient algorithms
Figure.
6. a kind of cloud detection device based on remote sensing image, which is characterized in that described device includes:
Binaryzation mark module obtains preliminary bipartite graph for carrying out binaryzation label to remote sensing image using threshold method；
Connected region identification module identifies to obtain three component of connected region for carrying out connected region to the preliminary bipartite graph；
Cloud detection module obtains the alpha figure of the remote sensing image for carrying out cloud detection to three component of connected region.
7. device as claimed in claim 6, which is characterized in that the binaryzation mark module includes:
Gray threshold computing unit, for calculating separately gray threshold of the remote sensing image under a variety of threshold methods；
Binaryzation marking unit, for being directed to each pixel of remote sensing image, according to the gray threshold under each threshold method to described
Pixel carries out binaryzation label；
Bipartite graph generation unit obtains preliminary bipartite graph for marking by the binaryzation of each pixel.
8. device as claimed in claim 6, which is characterized in that the connected region identification module includes:
Group's marking unit records each group for carrying out progressive scan group to the preliminary bipartite graph using Run- Length Coding
Starting point, terminal and line number；
Equivalence is to recording unit, for scanning continuous group, it is set in equivalent table initial markers and record it is of equal value right；
Marking unit is traversed, for traversing the label for starting group, the equivalent sequence in equivalent table is searched, it is same to assign equivalent sequence
Label；
Connected component labeling unit obtains connected component labeling for inserting the label of each group in the preliminary bipartite graph
Figure is set as determining prospect, obtains three components using the focus point of connected region as seed point.
9. a kind of electronic equipment, which is characterized in that the electronic equipment includes:
Processor；And
The memory being connect with the processor communication；Wherein,
The memory is stored with readable instruction, and the readable instruction realizes such as claim when being executed by the processor
The described in any item methods of 1-5.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer journey
Sequence realizes the method according to claim 1 to 5 when executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811536035.1A CN109584262A (en) | 2018-12-15 | 2018-12-15 | Cloud detection method of optic, device and electronic equipment based on remote sensing image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811536035.1A CN109584262A (en) | 2018-12-15 | 2018-12-15 | Cloud detection method of optic, device and electronic equipment based on remote sensing image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109584262A true CN109584262A (en) | 2019-04-05 |
Family
ID=65928227
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811536035.1A Pending CN109584262A (en) | 2018-12-15 | 2018-12-15 | Cloud detection method of optic, device and electronic equipment based on remote sensing image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109584262A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110503704A (en) * | 2019-08-27 | 2019-11-26 | 北京迈格威科技有限公司 | Building method, device and the electronic equipment of three components |
| CN110969609A (en) * | 2019-11-29 | 2020-04-07 | 华中科技大学 | Microwave remote sensing bright temperature image cloud detection method |
| CN113723381A (en) * | 2021-11-03 | 2021-11-30 | 航天宏图信息技术股份有限公司 | Cloud detection method, device, equipment and medium |
| CN115880129A (en) * | 2022-12-12 | 2023-03-31 | 华中光电技术研究所(中国船舶集团有限公司第七一七研究所) | High frame rate binary image connected domain automatic marking method and system based on FPGA |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1750046A (en) * | 2005-10-20 | 2006-03-22 | 浙江大学 | Three-dimensional ink and wash effect rendering method based on graphic processor |
| CN103034980A (en) * | 2012-12-07 | 2013-04-10 | 上海电机学院 | Self-adaptation morphological filtering system and method applied to calorescence correlation imaging |
| CN103196917A (en) * | 2013-03-13 | 2013-07-10 | 同济大学 | CCD linear array camera-based online rolled sheet material surface flaw detection system and detection method thereof |
| CN103235952A (en) * | 2013-04-03 | 2013-08-07 | 中国科学院遥感应用研究所 | Urban green space measuring method based on high-resolution remote sensing image |
| CN104484668A (en) * | 2015-01-19 | 2015-04-01 | 武汉大学 | Unmanned aerial vehicle multi-overlapped-remote-sensing-image method for extracting building contour line |
| CN104574448A (en) * | 2014-11-28 | 2015-04-29 | 浙江工商大学 | Method for identifying connected pixel blocks |
| CN105096326A (en) * | 2015-08-13 | 2015-11-25 | 丽水学院 | Laplace cutout matrix method by using moving least square method |
| CN105225236A (en) * | 2015-09-21 | 2016-01-06 | 中国科学院半导体研究所 | A kind of bianry image connected region paralleled detection method and system |
| CN107084991A (en) * | 2017-04-10 | 2017-08-22 | 武汉理工大学 | The detection of quartz pushrod bubble and quality grading method based on machine vision |
| CN108062759A (en) * | 2018-01-25 | 2018-05-22 | 华中科技大学 | A kind of more pixel-parallel labeling methods and system for being used to mark bianry image |
-
2018
- 2018-12-15 CN CN201811536035.1A patent/CN109584262A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1750046A (en) * | 2005-10-20 | 2006-03-22 | 浙江大学 | Three-dimensional ink and wash effect rendering method based on graphic processor |
| CN103034980A (en) * | 2012-12-07 | 2013-04-10 | 上海电机学院 | Self-adaptation morphological filtering system and method applied to calorescence correlation imaging |
| CN103196917A (en) * | 2013-03-13 | 2013-07-10 | 同济大学 | CCD linear array camera-based online rolled sheet material surface flaw detection system and detection method thereof |
| CN103235952A (en) * | 2013-04-03 | 2013-08-07 | 中国科学院遥感应用研究所 | Urban green space measuring method based on high-resolution remote sensing image |
| CN104574448A (en) * | 2014-11-28 | 2015-04-29 | 浙江工商大学 | Method for identifying connected pixel blocks |
| CN104484668A (en) * | 2015-01-19 | 2015-04-01 | 武汉大学 | Unmanned aerial vehicle multi-overlapped-remote-sensing-image method for extracting building contour line |
| CN105096326A (en) * | 2015-08-13 | 2015-11-25 | 丽水学院 | Laplace cutout matrix method by using moving least square method |
| CN105225236A (en) * | 2015-09-21 | 2016-01-06 | 中国科学院半导体研究所 | A kind of bianry image connected region paralleled detection method and system |
| CN107084991A (en) * | 2017-04-10 | 2017-08-22 | 武汉理工大学 | The detection of quartz pushrod bubble and quality grading method based on machine vision |
| CN108062759A (en) * | 2018-01-25 | 2018-05-22 | 华中科技大学 | A kind of more pixel-parallel labeling methods and system for being used to mark bianry image |
Non-Patent Citations (6)
| Title |
|---|
| J.YANG 等: "An automated cloud detection method based on the green channel of total-sky visible images", 《ATMOSPHERIC MEASUREMENT TECHNIQUES》 * |
| 周红妹 等: "气象卫星大雾遥感自动识别技术研究", 《热带气象学报》 * |
| 姜圆香 等: "基于连通域筛选和腐蚀重构的钢材裂痕检测", 《长江大学学报（自然版）》 * |
| 戴薇: "基于暗通道的遥感图像云检测算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 杨爽 等: "基于视觉显著性和区域生长的红外目标提取方法", 《沈阳航空航天大学学报》 * |
| 罗会兰 等: "一种基于集成学习技术的图像分割算法的研究", 《江西理工大学学报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110503704A (en) * | 2019-08-27 | 2019-11-26 | 北京迈格威科技有限公司 | Building method, device and the electronic equipment of three components |
| CN110969609A (en) * | 2019-11-29 | 2020-04-07 | 华中科技大学 | Microwave remote sensing bright temperature image cloud detection method |
| CN110969609B (en) * | 2019-11-29 | 2022-04-29 | 华中科技大学 | Microwave remote sensing bright temperature image cloud detection method |
| CN113723381A (en) * | 2021-11-03 | 2021-11-30 | 航天宏图信息技术股份有限公司 | Cloud detection method, device, equipment and medium |
| CN115880129A (en) * | 2022-12-12 | 2023-03-31 | 华中光电技术研究所(中国船舶集团有限公司第七一七研究所) | High frame rate binary image connected domain automatic marking method and system based on FPGA |
| CN115880129B (en) * | 2022-12-12 | 2024-06-04 | 华中光电技术研究所(中国船舶集团有限公司第七一七研究所) | Automatic marking method and system for high-frame-rate binary image connected domain based on FPGA |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022213879A1 (en) | Target object detection method and apparatus, and computer device and storage medium | |
| CN109961009B (en) | Pedestrian detection method, system, device and storage medium based on deep learning | |
| CN109584262A (en) | Cloud detection method of optic, device and electronic equipment based on remote sensing image | |
| US11392792B2 (en) | Method and apparatus for generating vehicle damage information | |
| US20210342999A1 (en) | System and method for training a damage identification model | |
| CN106874826A (en) | Face key point-tracking method and device | |
| CN111340126A (en) | Article identification method and device, computer equipment and storage medium | |
| CN110910445B (en) | Object size detection method, device, detection equipment and storage medium | |
| CN112862757A (en) | Weight evaluation system based on computer vision technology and implementation method | |
| CN115797846A (en) | Wind power generation blade block defect comparison method and device and electronic equipment | |
| CN110657760B (en) | Method and device for measuring space area based on artificial intelligence and storage medium | |
| CN114332602A (en) | Commodity identification method of intelligent container | |
| CN112580584A (en) | Method, device and system for detecting standing behavior and storage medium | |
| CN111523423A (en) | Power equipment identification method and device | |
| CN113269730B (en) | Image processing method, image processing device, computer equipment and storage medium | |
| CN115063473A (en) | Object height detection method and device, computer equipment and storage medium | |
| CN102999763B (en) | Based on the top-down vision significance extracting method of scale selection | |
| CN115810011A (en) | Training method, device and equipment for anomaly detection network and anomaly detection method, device and equipment | |
| CN110399900A (en) | Method for checking object, device, equipment and medium | |
| CN115424000A (en) | Pointer instrument identification method, system, equipment and storage medium | |
| CN110705604B (en) | AI algorithm-based dynamic model detection method | |
| CN109116746A (en) | A kind of smart home system | |
| CN115965856B (en) | Image detection model construction method, device, computer equipment and storage medium | |
| CN116612474B (en) | Object detection method, device, computer equipment and computer readable storage medium | |
| CN117495853B (en) | Video data processing method, device and storage medium |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190405 |
|
| RJ01 | Rejection of invention patent application after publication |