CN103714339A - SAR image road damaging information extracting method based on vector data - Google Patents
SAR image road damaging information extracting method based on vector data Download PDFInfo
- Publication number
- CN103714339A CN103714339A CN201310750917.9A CN201310750917A CN103714339A CN 103714339 A CN103714339 A CN 103714339A CN 201310750917 A CN201310750917 A CN 201310750917A CN 103714339 A CN103714339 A CN 103714339A
- Authority
- CN
- China
- Prior art keywords
- road
- vector data
- sar image
- width
- doubtful
- 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
Images
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
- G06V20/182—Network patterns, e.g. roads or rivers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
- G06F18/24155—Bayesian classification
-
- 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
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
Provided is an SAR image road damaging information extracting method based on vector data. According to the range of SAR images after a disaster, vector data of a corresponding area are obtained; the vector data are projected to a coordinate system of the SAR images and then are registered on the SAR images; a suspected road damaging area of the SAR images is extracted, wherein a road detecting operator is used for line detecting, road width information and the road vector data are subjected to shape level set segmentation, and the road damaging area is obtained by fusion; and a Bayes network model is established to carry out further judging on the suspected road damaging area, and road damaging information is extracted. According to the method, the vector data are used as prior information for helping SAR image road changing detecting, detecting rate is high, a breaking zone can be well extracted through a method with line detecting and a shape level set being combined, the omission factor is low, interference information can be effectively removed through the established Bayes network model, and false alarm in road damaging extracting is lowered.
Description
Technical field
The present invention relates to Remote Sensing Image Processing Technology field, particularly the High-resolution SAR Images road damage information extracting method of a kind of vector data under auxiliary.
Background technology
Road is national economy and military artery, on military and civilian, all has very important significance.When various disasters occur, road lifeline just may be blocked, for example flood, landslide, the disasteies such as rubble flow all may cause the obstruction of road, make to send rescue personnel and toward disaster area, transport rescue material to be subject to great obstruction, to rescue and relief work, bring huge inconvenience.After disaster occurs, because the mode of artificial field exploring is the work taking time and effort, remote sensing technology is because the feature of " day eye " makes it to become the very important mode of road damage of extracting.
Make a general survey of the various countries scholar research that damage is extracted to remote sensing image road in recent years, Most scholars all uses multispectral optical image to carry out road damage extraction as data source, and the disaster-stricken rear boisterous impact of the quick obtaining of optical image is larger, it is limited by very large in rescue and relief work decision-making.SAR is due to its special imaging mechanism, can overcome the impact of weather and illumination condition, round-the-clock, round-the-clock, observation are on a large scale carried out in target area, therefore after disaster occurs, utilize SAR image to extract road, find out road damage region and have more advantage.Although the method extracted of SAR image road that had many scholar's research, rarely has people to set foot in for the road damage extraction of SAR image, especially High Resolution SAR Images.
Summary of the invention
For the problems referred to above, the present invention adopts vector data to extract as the damage of aid in guide road, and the road damage of having realized High Resolution SAR Images extracts.
Technical scheme of the present invention is a kind of SAR image road damage information extracting method based on vector data, comprises the following steps:
Step 1, according to the scope of the SAR image after calamity, obtains the vector data of corresponding region, and described vector data comprises road vectors data;
Step 2, after the vector data of step 1 gained corresponding region is projected to the coordinate system of SAR image, is registrated to vector data on SAR image;
Step 3, the doubtful road damage district that extracts SAR image, comprises following sub-step,
Step 3.1, in the buffer zone of setting up according to road vectors data, utilize Road Detection operator to carry out line detection, obtain Road segment base unit and road width information, the position of finding road fracture according to the testing result of Road segment base unit, obtains doubtful road damage district;
Step 3.2, step 3.1 gained road width information is combined with road vectors data, obtains the prior shape constraint that shape level set is cut apart, utilize prior shape constraint to carry out level set movements, obtain the segmentation result of road area and find the fracture position of road, obtain doubtful road damage district;
Step 3.3, merges step 3.1 and the doubtful road damage of 3.2 gained district, obtains final doubtful road damage district;
Step 4, sets up the doubtful road damage district that Bayesian network model extracts step 3 and further judges, extracts road damage information.
And, in step 1, from OpenStreetMap server, download vector data.
And in step 1, described vector data also comprises outline of house vector data and waters contour vector data.
And step 1 comprises following sub-step,
Step 1.1, projects to 4 angular coordinates of SAR image under the coordinate system of vector data, establishes projection gained latitude and longitude coordinates and is respectively (X
1, Y
1), (X
2, Y
2), (X
3, Y
3), (X
4, Y
4);
Step 1.2, asks for X
1, X
2, X
3, X
4maximum, minimum value X
max, X
minand Y
1, Y
2, Y
3, Y
4maximum, minimum value Y
max, Y
min;
Step 1.3, by (X
min, Y
max), (X
min, Y
min), (X
max, Y
max), (X
max, Y
min) as 4 angle points of the scope of download, obtain the vector data of corresponding region.
And in step 3.1, described Road Detection operator is as follows,
The template of Road Detection operator is that width is the rectangle that 2W, length are L, and the middle section of template is that width is the rectangle that W, length are L, and it is the rectangle that W/2, length are L that the region, the left and right sides of template is respectively width; Adopt formwork calculation response Gap=min(1-avr2/avr1,1-avr2/avr3), if calculate Gap<0, make Gap=0.
And, in step 3.1, utilize Road Detection operator to carry out line detection, obtain Road segment base unit and road width information, implementation is,
If certain width SAR image resolution is N rice, make W
1=8/N, W
2=16/N, W
3=24/N, W
4=32/N, establishes road width W=W successively to every section of road vectors in road vectors data
1, W
2, W
3, W
4, under each value, carry out following operation, this section of road vectors both sides, expand, set up the buffer zone that overall width is 4W, mobile Road Detection operator detects in buffer zone, records this line segment when response Gap is greater than predetermined threshold value; The move mode of mobile Road Detection operator is, from this section of road vectors starting point, buffer zone is divided into several minibuffers district, and the template length of each minibuffer section length and mobile Road Detection operator is consistent; In each minibuffer district, from road vectors center along perpendicular to road vectors direction toward two side shiftings, the width W/4 of being often separated by calculates once; Record template center's line segment that each surpasses corresponding mobile Road Detection operator under the result of calculation of threshold value, and calculate from detected line segment sum in buffer zone;
Relatively under each value from detected line segment sum in buffer zone, the developed width that the value of selecting to detect maximum line segments is road.
And, in step 3.2, to every section of road vectors in road vectors data, using this section of road vectors as center line, according to the width of road, toward both sides, be extended to a closed long strip type region as the shape constraining information of road, the length in long strip type region is the length of certain section of road, the width that width is road.
And in step 4, described Bayesian network model comprises 6 priori evidence information variable A, B, C, D, E, F, 2 observed reading G, H and doubtful road damage district actual attribute X, 6 priori evidence information variable A, B, C, D, E, F is the condition of doubtful road damage district actual attribute X, doubtful road damage district actual attribute X is 2 observed reading G, the condition of H; By solving the posterior probability of the actual attribute in doubtful road damage district, distribute as follows, the situation of selection maximum probability is as final decision result,
Wherein, P (X/A, B, C, D, E, F) is illustrated under various evidence conditions, and doubtful road damage district belongs to the prior probability of certain situation, P (G/X), and P (H/X) represents respectively the attribute in doubtful road damage district and the relation between observed reading.
And, described 6 priori evidence information variable A, B, C, D, E, F is respectively outline of house vector data, landslide point data, calamity kind and its intensity, DSM data, road vectors data, waters contour vector data; 2 observed reading G, H is respectively doubtful road damage district gray scale, doubtful road damage district texture.
The present invention proposes and a kind ofly improve that buffer zone line detects and level set is cut apart the method combining, extract the variation of SAR image road, and then set up Bayesian network model in conjunction with the observed reading of other aucillary documents and breakdown zone, these region of variation are further judged, reject false detection, extract the true damage information of road.
Accompanying drawing explanation
Fig. 1 is the general flow chart of the embodiment of the present invention.
Fig. 2 is the window model figure of the Road Detection operator of the embodiment of the present invention.
Fig. 3 is the Road Detection template of the embodiment of the present invention move mode figure in buffer zone.
Fig. 4 is the combination breakdown zone observed reading of the embodiment of the present invention and the Bayesian network model figure of aucillary document.
Embodiment
The invention provides the auxiliary lower SAR image road of a kind of geographical vector data and damage extracting method.Mainly with High-resolution SAR Images road damage, be extracted as research contents, adopt geographical vector data as prior imformation, based on improved buffer zone line, detect and level set is cut apart integrated method and found doubtful road damage district.Further consider in SAR image foldedly cover, the interference of coherent spot and the factors such as complicacy of road background itself, the present invention takes Bayes posterior probability model to carry out depth analysis to doubtful road damage district, and then extraction road damage.
Below in conjunction with drawings and Examples, describe technical solution of the present invention in detail.
OpenStreetMap (being called for short OSM) is a Internet map cooperation plan of increasing income, user can be free download various vector datas (comprising road vectors data, house and water body contour vector data etc.) from the Internet, its real-time property is strong and precision is high, can be well as prior imformation service road damage extraction.Embodiment adopts disclosed geographical vector data OpenStreetMap auxiliary, and High-resolution SAR Images road damage extracting method is provided.
Technical solution of the present invention can adopt computer software technology to realize operation automatically.As shown in Figure 1, the technical scheme flow process of embodiment comprises the following steps:
Obtaining of step 1 vector data.According to SAR image capturing range can automatic acquisition up-to-date OSM road vectors data, outline of house vector data, the waters contour vector data of corresponding scope, every kind of vector data in SAR image capturing range may comprise respectively multistage vector line segment, and every section of vector line segment is comprised of a series of point.Wherein road vectors data are used as the front data source of calamity, the change information of assisted extraction road; Latter two vector data can be used as aucillary document in order to being whether damage judgement for further analysis to change information.During concrete enforcement, can, according to the scope of post-disaster high-resolution SAR image, automatically download from network the vector data of corresponding region.
Embodiment is used OpenStreetMap API (XAPI), selects to download region, builds a range boundary frame, then builds a download address, from OpenStreetMap server, downloads vector raw data.Definite mode of download scope is:
(1) 4 angular coordinates of SAR image are projected under the coordinate system of OpenStreetMap vector data, establish projection gained latitude and longitude coordinates and be respectively (X
1, Y
1), (X
2, Y
2), (X
3, Y
3), (X
4, Y
4).
(2) ask for X
1, X
2, X
3, X
4maximum, minimum value X
max, X
minand Y
1, Y
2, Y
3, Y
4maximum, minimum value Y
max, Y
min.
(3) by (X
min, Y
max), (X
min, Y
min), (X
max, Y
max), (X
max, Y
min) as 4 angle points of the scope of download, obtain the vector data of corresponding region.
In practical operation situation, owing to existing certain deviation in vector data and SAR image coordinate, the scope reality of therefore downloading can be slightly larger than the determined scope of said method.
Step 2 is registrated to vector data on SAR image after the vector data of corresponding region is projected to the coordinate system of SAR image.
After obtaining OpenStreetMap vector data and being projected to SAR coordinate systems in image, between SAR image and vector data, may there is grid deviation in embodiment.During concrete enforcement, can SAR image as benchmark, registration mode with reference to same place in prior art or line of the same name is corrected OpenStreetMap vector data, comprises step 1 gained road vectors data, outline of house vector data, waters contour vector data are all carried out to registration.
If there is the obvious point of crossing of large measure feature and flex point on image, there is point of crossing or flex point in the road on road vectors data and respective image for example, take so to select the registration mode of same place comparatively suitable, on vector data and SAR image, select respectively respective quadrature crunode or flex point as same place to rear, the method that can utilize polynomial expression in prior art to correct calculates the coefficient of conversion.Then to each point on each section of vector line segment, according to polynomial equation and the conversion coefficient that calculated, calculate the position after correcting.For the purpose of raising the efficiency, while specifically implementing, can for example, by human-computer interaction interface (touch-screen), to user, provide vector data and SAR image, same place can be specified selection by user.
Consider the situation that some image flex point and point of crossing are difficult to seek, can adopt the mode of line of the same name, only need to select the line segment that is located on the same line on vector data and SAR image, do not need between line segment strictly coupling.In order to simplify user's operation, can the design software method of operation be, on vector data, user only needs reconnaissance, software is all vector line segments of traversal automatically, and find out from this and put nearest line segment (2 straight-line segments that are linked to be on vector) as the line segment being selected.Select the concrete steps suggestion of line segment registration as follows:
(1) on SAR image, select several uneven line segments.In order to guarantee the precision of registration, the angle of choosing between line segment can not be too little.
(2) on vector data, select corresponding vector line segment, practising way is, when the nearest point of the corresponding line segment of user's chosen distance, system is all vector line segments of traversal automatically, choose to comprise this point (this spot projection be positioned at line segment to the vertical point of line segment within) and apart from minimum vector line segment.
(3) according to the line segment of selecting on SAR image and vector data, ask for respectively intersection point of line segments.
(4), using the point of crossing of asking in step (3) as same place, utilize the method for choosing same place registration of introducing that vector data is registrated on image above.The method that can utilize polynomial expression in prior art to correct calculates the coefficient of conversion, then to each point on each section of vector, according to polynomial equation and the conversion coefficient that calculated, calculates and corrects position afterwards.
According to the image after registration and vector, carry out subsequent step.
The method that step 3 joint line detects and shape level set is cut apart is extracted the doubtful road damage district of SAR image.
Embodiment implementation is as follows,
(1) under the OpenStreetMap road vectors data after step 2 registration auxiliary, in the buffer zone of setting up according to road vectors data, utilize Road Detection operator to carry out line detection, can obtain line segment primitive and the road width information of road.Based on breakdown zone, can not detect the thought of road-center line primitives, can according to the testing result of Road segment base unit, find the position of road fracture, obtain a part of doubtful road damage district.
(2) the road width information that the present invention obtains line detection is combined with road vectors data, the prior shape constraint of cutting apart as shape level set.Utilize prior shape constraint to carry out level set movements, also can obtain the segmentation result of road area and find the fracture position of road, also obtain a part of doubtful road damage district.
(3) doubtful road damage district in line testing result and shape level set segmentation result is merged, to reduce the generation of undetected situation.
In (1), the method that obtains road width is based on following thought: when the middle section width of Road Detection template more caters to the developed width of certain section of road, utilizing so this template in this section of buffer zone, to detect the sum of the Road segment base unit obtaining will be more.
In (2), the method of setting up the constraint of shape level set prior shape is: using each section of road vectors line as center line, according to the width of road, toward both sides, be extended to a closed long strip type region, the length that the length in long strip type region is certain section of road, the width that width is road.
Embodiment, under OpenStrertMap road vectors data auxiliary, utilizes the Road Detection operator that improves D1 operator to detect road-center line primitives and road width information.The present invention improves the template of the Road Detection operator adopting after D1 operator and sees accompanying drawing 2.Wherein template width is 2W, and middle section width is that W(W is the road width of selecting, and is selectable variable); L is template length, and template is that width is the rectangle that 2W, length are L, and it is the rectangle that W, length are L that the center line segment that in template, length is L can be divided into template two width.The operator of embodiment is mainly considered central area and region, left and right, and the middle section of template is that width is the rectangle that W, length are L, and it is the rectangle that W/2, length are L that the region, the left and right sides of template is respectively width.During concrete enforcement, L can calculate according to road vectors line segment length, for example certain section of road vectors length along path is S, general default L is 2 times of template width W, this section of road vectors line segment can be divided into N=integer(S/L so) integer section, the length correction of L is L=S/N the most at last, and integer represents to round.If its middle section (as black region part in Fig. 2) gray average is avr2, left and right sides area grayscale average is respectively avr1, avr3.Adopt the method for the response Gap of this line segment of formwork calculation to be:
Gap=min（1-avr2/avr1,1-avr2/avr3）；
If above formula calculates Gap<0, make Gap=0, otherwise preservation above formula result of calculation is constant.
Road is divided into various ranks according to the difference of its function, and the road width of different stage is different.The road in 2 tracks, 4 tracks, 6 tracks, 8 tracks is that we are modal, and due to the impact on the other walkway of road, the developed width that road presents on High-resolution SAR Images also outline is larger.What the embodiment of the present invention was rough is divided into 8 meters road according to width, and 16 meters, 24 meters, 32 meters of several ranks, suppose that certain width SAR image resolution is N rice, and the selectable width of road on SAR image is respectively W so
1=8/N, W
2=16/N, W
3=24/N, W
4=32/N.During concrete enforcement, can be according to real road situation degree of establishment.
The method that embodiment utilizes line to detect, every section of road vectors in road vectors data is processed and obtained Road segment base concrete grammar first and road width and be:
(1) first suppose that road width W is W
1, corresponding Road Detection operator width is 2W
1, the width of middle section is W
1, then certain section of road vectors both sides, expand, setting up overall width is 4W
1buffer zone.Mobile Road Detection operator (seeing above-mentioned improvement D1 operator and next Road Detection operator) detects in buffer zone, when being greater than predetermined threshold value (those skilled in the art can preset value voluntarily, and suggestion value is between 0.12-0.2), records operator response value Gap this line segment.The move mode that detects operator is: at width, be 4W
1buffer zone in, Road Detection template, from this section of road vectors starting point, is divided into buffer area the segment buffer area of several sections (being N section), each segmentation zonule is designated as Yi Ge minibuffer district, length is L.In each minibuffer district, from road vectors center along perpendicular to road vectors direction toward two side shiftings, the width W of being often separated by
1/ 4 calculate once, establish road vectors center and be designated as 0, by template from 0 toward a side shifting to W
1/ 4, W
1/ 2,3W
1/ 4, W
1, 5W
1/ 4,3W
1/ 2,7W
1/ 4,2W
1place is calculated respectively, in the same movement of opposite side calculating respectively.Having calculated for 17 times of Yi Ge minibuffer district moves down the length L of a template afterwards along buffer zone, continue to repeat aforesaid operations, until the detection of whole buffer zone is complete in the minibuffer district of next part.Record each template center's line segment that surpasses corresponding mobile Road Detection operator under the result of calculation of threshold value and calculate from detected line segment sum in buffer zone.The move mode of Road Detection template in buffer zone as shown in Figure 3.
(2) road width is changed into W
2, W
3, W
4re-establish again respectively buffer zone and with the Road Detection operator that the present invention proposes, calculate the response of each position operator, and recording responses value surpasses the line segment of threshold value.
(3), if the middle section width of Road Detection operator more agrees with the developed width of road, in buffer zone, detect so satisfactory line segment also just more.Therefore comparison operator middle section width is respectively W
1, W
2, W
3, W
4under different situations from detected line segment number in buffer zone, selection detects the developed width that middle section width that line segment number is maximum is road, and using detecting specifically the primary data of the testing result of maximum line segments as subsequent treatment, current detected each line segment over threshold value of take is road primitives.
After above-mentioned road buffering district line detecting step, obtained line segment primitive and the road width information of each section of road vectors.Owing between Mei Duan minibuffer district being (the seeing shown in accompanying drawing 2) not exclusively overlapping, even if therefore road ruptures, between road primitives, also may there is trickle fracture.During concrete enforcement, after detecting road primitives, can utilize the information such as distance between line segment primitive, curvature, curvature variation, road primitives is organized into groups and become broken line.For example, because road both sides may exist the interfere information similar with road (shade similar with roadway characteristic that in a row buildings forms), therefore after marshalling, also to screen marshalling result in addition.The algorithm of screening is that the thought that is weaker than real road information based on interfere information is carried out: for the broken line after marshalling, if two broken line overlapping range of projection in road vectors reaches certain threshold value (can be made as shorter broken line 1/3), the Road segment base unit hop count comprising according to broken line, reject the broken line of line segment primitive negligible amounts, retain the larger broken line of line segment primitive quantity.After marshalling and screening step, the broken line obtaining is projected on road vectors, wherein there is not the road vectors scope of line segment projection, be identified as the region of fracture.
Through buffer zone line, detect the road width information obtaining and combine with road vectors data, the prior shape constraint of formed shape level set.Concrete grammar is to every section of road vectors in road vectors data, using this section of road vectors as center line, according to the width of road, toward both sides, be extended to a closed long strip type region as the shape constraining information of road, the length in long strip type region is the length of certain section of road, the width that width is road.After obtaining the shape constraining information of road, utilize the method that in prior art, shape level set is cut apart to cut apart the road area that obtains every section of road vectors.To the road area of every section of road vectors, can carry out segmentation judgement, for example according to L, carry out segmentation, and according to every segment road area institute occupied road number of pixels number judge whether to rupture (for example, when every segment road area institute occupied road number of pixels is less than corresponding predetermined threshold value, is judged as and ruptures), thereby find out the region of road fracture.The present invention adds the shape constraining of road target, comes bound level collection just to cut apart can effectively reject interference by the prior shape of target, obtains and extracts more accurately result.
In order to reduce may exist undetected, the present invention doubtful road damage district that two kinds of methods based on line detects and shape level set is cut apart extract by this combines, and both results is merged to (can adopt and ask mode also to merge).
Step 4 is utilized various supplementarys and doubtful road damage district observed reading, sets up Bayesian network model, and the doubtful road damage district that step 3 is extracted further judges, extracts the damage information of road.
In traditional road damage extracting method, just according to image after calamity before calamity (image before calamity can utilize road vectors data to replace), extract road change information, then the damage region using the region of these variations as road.But in fact due to the impact of other various interference, it must be not that true damage is caused that variation detects the road change information obtaining, therefore the present invention has adopted Bayesian network model to be further analyzed judgement to the breakdown zone of detecting, and has higher fiduciary level.
The present invention is divided into following a few class situation the doubtful damage of the SAR image region, road extracting above:
(1) road both sides buildings collapses after earthquake, causes road congestion.
(2) the landslide blocked path that earthquake or other reasons cause.
(3) reason such as heavy rain or seismic checked-up lake causes road to be inundated with flood.
(4) folded the covering that high-lager building and topographic relief cause.
(5) complicated intersection and bridge.
(6) coherent speckle noise and other interference.
What be associated with this six events above comprises various aucillary documents and the observed reading of these events on image that determines its prior probability.First relevant with three damages is above exactly concrete calamity kind, such as house collapse is substantially all due to seismic, if there is not earthquake, can think so that because house collapse causes the probability of road damage be 0, if there is earthquake, the probability of house collapse is also relevant with concrete earthquake magnitude so.Also have in addition six events of a lot of reasons and this relevant, will be illustrated respectively below.
House collapse is mainly with whether earthquake to occur relevant with its earthquake magnitude, but the road damage that house collapse causes is with the Range-based of house and road.If do not exist house this section of road both sides, this section of region is just 0 because house collapse causes the probability of damage so.Therefore the distance of house and road is one of reason causing this class damage.
The generation of landslide is also relevant with the geologic condition in this region, if there are landslide point data, so just landslide can be put to data and this place and landslide occurs cause the probability of road damage to connect.The gradient in the other mountain region of this section of road etc. is also to cause the probability of road damage to link together with landslide in addition, if judge road both sides according to DEM/DSM, there is no hillside, belongs to level land, so also just can not landslide occur.
Flood inundation on tracks is relevant with this section of residing elevation of road, the DEM/DSM model on this ground and its important contacting that whether can be inundated with flood.The other waters information in this section of region is in addition also that the probability that is inundated with flood with this section of road is relevant.
Except disaster causes real roads damage, folded covering by high-lager building and topographic relief caused, and it is to be mainly associated with DEM/DSM information; The fracture that road causes through waters (bridge) is relevant to waters information.
The actual attribute of breakdown zone, except being associated with these six evidences, is also direct correlation with this observed reading showing on SAR image of this region in addition.The observed reading that the road that different situations causes breaks on SAR image is distinguished to some extent, and generally observed reading is represented with gray scale and texture, and accompanying drawing 4 is Bayesian network models of the road damage taked of the present invention.
Fundamental purpose of the present invention is in conjunction with actual observed value judgement breakdown zone, to belong to various situation possibilities in order to calculate under various evidences auxiliary, so do not need to calculate the joint probability density between all variablees.Except outline of house vector data, waters contour vector data, the present invention also can consider other aucillary documents.6 priori evidence information variables that expose in network model shown in accompanying drawing 4---outline of house data (being the outline of house vector data after step 2 registration), landslide point data, calamity kind and its intensity, DSM data, road vectors data, waters information (being the waters contour vector data after step 2 registration) is used respectively A, B, C, D, E, F replaces, 2 observed readings---doubtful road damage district gray scale, doubtful road damage district texture is used respectively G, H represents, the doubtful road damage district actual attribute of hiding represents with X, 6 priori evidence information variable A, B, C, D, E, F is the condition of doubtful road damage district actual attribute X, doubtful road damage district actual attribute X is 2 observed reading G, the condition of H.The conditional probability that the present invention need to ask for is so P (X/A, B, C, D, E, F, G, H), and its computing formula is:
Wherein P (X/A, B, C, D, E, F) is illustrated under various evidence conditions, and this doubtful road damage district belongs to the prior probability of certain situation.This conditional probability is determined by expertise, in the present invention, as known conditions, pre-enters.
By A, B, C, D, E, these six aucillary documents of F just can obtain the prior probability P (X/A, B, C, D, E, F) of doubtful road damage district attribute.Calamity kind generally has earthquake, heavy rain etc., the reason of the road damage causing has landslide, house collapse, flood etc., those skilled in the art can preset respective function as the case may be voluntarily, for example X belongs to the probability P (X=Rockslid/A of the road damage that landslide (Rockslid) causes, B, C, D, E, F) large I is set as to disaster intensity (C determines by variable) with by D(DEM/DSM data) terrain slope that determines is directly proportional, if A belongs to landslide point its probability should be larger so simultaneously; X belongs to the probability P (X=BuildingCollaps/A of the road damage that house collapse (BuildingCollaps) causes, B, C, D, E, F) large I is set as being directly proportional to disaster intensity (C determines by variable), and by A(outline of house data) breakdown zone that determines and the distance in house be inversely proportional to; X belongs to probability P (X=Flood/A, B, the C that flood (Flood) floods, D, E, F) large I be set as being directly proportional to disaster intensity (C determines by variable), and by F(waters information) breakdown zone determining is with the distance of water body with by D(DEM/DSM data) elevation of decision is inversely proportional to.To belong to the probability of certain situation all relevant to calamity kind (C determines by variable) for X in addition, and the disasteies such as earthquake heavy rain cause that the probability of different kinds of roads damage is also different.
Other factors also can have influence on the probability that X belongs to various true damages, such as when according to road vectors data E, judgement learns that this breakdown zone is positioned on intersection, the probability that breakdown zone belongs to true damage will reduce, and the probability that belongs to interfere information increases (because intersection often can be detected as breakdown zone in the present invention's the 3rd step); When according to road vectors data E and waters data F, judgement learns that this breakdown zone is positioned on bridge, the probability that breakdown zone belongs to true damage will reduce, the probability that belongs to interfere information increases (because when bridge is parallel with radar incident direction, owing to there is no dihedral angle reflection effect, and the reflection strength of water body and road is all very low, may cause the bridge invisible on SAR image); The folded scope of covering of asking for according to the sensor parameters of DEM/DSM and SAR image, judges breakdown zone and is positioned at foldedly when covering scope, and to belong to the probability of true damage be just 0 in breakdown zone so.
P (G/X), P (H/X) represents respectively the attribute in doubtful road damage district and the relation between its observed reading, can calculate this two probability by solving respectively when X belongs to the distribution of observed reading under different situations, during concrete enforcement, initial probability distribution can be made as to normal distribution, then with observed reading training, obtain its parameter.By priori, can determine that X belongs to the characteristic (function that comprises unknown parameter) of its gray scale and texture under different situations, if there is a large amount of observed data, can solve unknown parameter by the method for maximum likelihood method, thereby obtain, belong to the probability distribution in the situation of certain attribute as X.For example, if X equals the road damage that causes of landslide, so just need to utilize landslide to cause that a large amount of observed readings in the region of road damage train, draw the gray scale of road damage and the probability distribution of texture that come down and cause.
Pre-determine P (X/A, B, C, D, E, F) and P (G/X), P (H/X), just can solve in the situation based on various evidences and observed reading, the posterior probability distribution P (X/A of the actual attribute in doubtful road damage district, B, C, D, E, F, G, H), select the situation of maximum probability as final decision result.
Above content is in conjunction with concrete embodiment further description made for the present invention, can not assert that specific embodiment of the invention is confined to these explanations.For general technical staff of the technical field of the invention, without departing from the inventive concept of the premise, can also make some simple deduction or replace, all should be considered as belonging to protection scope of the present invention.
Claims (9)
1. the damage of the SAR image road based on a vector data information extracting method, is characterized in that, comprises the following steps:
Step 1, according to the scope of the SAR image after calamity, obtains the vector data of corresponding region, and described vector data comprises road vectors data;
Step 2, after the vector data of step 1 gained corresponding region is projected to the coordinate system of SAR image, is registrated to vector data on SAR image;
Step 3, the doubtful road damage district that extracts SAR image, comprises following sub-step,
Step 3.1, in the buffer zone of setting up according to road vectors data, utilize Road Detection operator to carry out line detection, obtain Road segment base unit and road width information, the position of finding road fracture according to the testing result of Road segment base unit, obtains doubtful road damage district;
Step 3.2, step 3.1 gained road width information is combined with road vectors data, obtains the prior shape constraint that shape level set is cut apart, utilize prior shape constraint to carry out level set movements, obtain the segmentation result of road area and find the fracture position of road, obtain doubtful road damage district;
Step 3.3, merges step 3.1 and the doubtful road damage of 3.2 gained district, obtains final doubtful road damage district;
Step 4, sets up the doubtful road damage district that Bayesian network model extracts step 3 and further judges, extracts road damage information.
2. the SAR image road based on vector data is damaged information extracting method according to claim 1, it is characterized in that: in step 1, from OpenStreetMap server, download vector data.
3. according to the damage of the SAR image road based on vector data information extracting method described in claim 1 or 2, it is characterized in that: in step 1, described vector data also comprises outline of house vector data and waters contour vector data.
4. the SAR image road based on vector data is damaged information extracting method according to claim 3, it is characterized in that: step 1 comprises following sub-step,
Step 1.1, projects to 4 angular coordinates of SAR image under the coordinate system of vector data, establishes projection gained latitude and longitude coordinates and is respectively (X
1, Y
1), (X
2, Y
2), (X
3, Y
3), (X
4, Y
4);
Step 1.2, asks for X
1, X
2, X
3, X
4maximum, minimum value X
max, X
minand Y
1, Y
2, Y
3, Y
4maximum, minimum value Y
max, Y
min;
Step 1.3, by (X
min, Y
max), (X
min, Y
min), (X
max, Y
max), (X
max, Y
min) as 4 angle points of the scope of download, obtain the vector data of corresponding region.
5. the SAR image road based on vector data is damaged information extracting method according to claim 4, it is characterized in that: in step 3.1, described Road Detection operator is as follows,
The template of Road Detection operator is that width is the rectangle that 2W, length are L, and the middle section of template is that width is the rectangle that W, length are L, and it is the rectangle that W/2, length are L that the region, the left and right sides of template is respectively width; Adopt formwork calculation response Gap=min(1-avr2/avr1,1-avr2/avr3), if calculate Gap<0, make Gap=0.
6. the SAR image road based on vector data damage information extracting method according to claim 5, is characterized in that: in step 3.1, utilize Road Detection operator to carry out line detection, obtain Road segment base unit and road width information, implementation is,
If certain width SAR image resolution is N rice, make W
1=8/N, W
2=16/N, W
3=24/N, W
4=32/N, establishes road width W=W successively to every section of road vectors in road vectors data
1, W
2, W
3, W
4, under each value, carry out following operation, this section of road vectors both sides, expand, set up the buffer zone that overall width is 4W, mobile Road Detection operator detects in buffer zone, records this line segment when response Gap is greater than predetermined threshold value; The move mode of mobile Road Detection operator is, from this section of road vectors starting point, buffer area is divided into several minibuffers district, and the template length of each minibuffer section length and mobile Road Detection operator is consistent; In each minibuffer district, from road vectors center along perpendicular to road vectors direction toward two side shiftings, the width W/4 of being often separated by calculates once; Record template center's line segment that each surpasses corresponding mobile Road Detection operator under the result of calculation of threshold value, and calculate from detected line segment sum in buffer zone;
Relatively under each value from detected line segment sum in buffer zone, the developed width that the value of selecting to detect maximum line segments is road.
7. the SAR image road based on vector data is damaged information extracting method according to claim 6, it is characterized in that: in step 3.2, to every section of road vectors in road vectors data, using this section of road vectors as center line, according to the width of road, toward both sides, be extended to a closed long strip type region as the shape constraining information of road, the length in long strip type region is the length of certain section of road, the width that width is road.
8. the SAR image road based on vector data is damaged information extracting method according to claim 7, it is characterized in that: in step 4, described Bayesian network model comprises 6 priori evidence information variable A, B, C, D, E, F, 2 observed reading G, H and doubtful road damage district actual attribute X, 6 priori evidence information variable A, B, C, D, E, F is the condition of doubtful road damage district actual attribute X, doubtful road damage district actual attribute X is 2 observed reading G, the condition of H; By solving the posterior probability of the actual attribute in doubtful road damage district, distribute as follows, the situation of selection maximum probability is as final decision result,
Wherein, P (X/A, B, C, D, E, F) is illustrated under various evidence conditions, and doubtful road damage district belongs to the prior probability of certain situation, P (G/X), and P (H/X) represents respectively the attribute in doubtful road damage district and the relation between observed reading.
9. the SAR image road based on vector data is damaged information extracting method according to claim 8, it is characterized in that: described 6 priori evidence information variable A, B, C, D, E, F is respectively outline of house vector data, landslide point data, calamity kind and its intensity, DSM data, road vectors data, waters contour vector data; 2 observed reading G, H is respectively doubtful road damage district gray scale, doubtful road damage district texture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310750917.9A CN103714339B (en) | 2013-12-30 | 2013-12-30 | SAR image road damaging information extracting method based on vector data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310750917.9A CN103714339B (en) | 2013-12-30 | 2013-12-30 | SAR image road damaging information extracting method based on vector data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103714339A true CN103714339A (en) | 2014-04-09 |
| CN103714339B CN103714339B (en) | 2017-01-18 |
Family
ID=50407296
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310750917.9A Active CN103714339B (en) | 2013-12-30 | 2013-12-30 | SAR image road damaging information extracting method based on vector data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103714339B (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104050473A (en) * | 2014-05-20 | 2014-09-17 | 中国人民解放军理工大学 | Road data extraction method based on rectangular neighborhood analysis |
| CN104091369A (en) * | 2014-07-23 | 2014-10-08 | 武汉大学 | Unmanned aerial vehicle remote-sensing image building three-dimensional damage detection method |
| CN104268879A (en) * | 2014-09-28 | 2015-01-07 | 民政部国家减灾中心 | Physical building quantity damage evaluation method based on remote sensing multi-spectral images |
| CN104408437A (en) * | 2014-12-08 | 2015-03-11 | 电子科技大学 | Road detecting method based on synthetic aperture radar |
| CN106169086A (en) * | 2016-07-21 | 2016-11-30 | 武汉大学 | High-resolution optical image damage method for extracting roads under navigation data auxiliary |
| CN106295699A (en) * | 2016-08-11 | 2017-01-04 | 北京师范大学 | A kind of earthquake Damage assessment method and apparatus based on high-definition remote sensing data |
| CN107886505A (en) * | 2017-11-08 | 2018-04-06 | 电子科技大学 | A kind of synthetic aperture radar airfield detection method based on line primitives polymerization |
| CN108229273A (en) * | 2017-02-27 | 2018-06-29 | 北京市商汤科技开发有限公司 | Multilayer neural network model training, the method and apparatus of roadway characteristic identification |
| CN108388828A (en) * | 2017-07-13 | 2018-08-10 | 中国科学院遥感与数字地球研究所 | A kind of seashore wetland land cover pattern information extracting method of comprehensive multi- source Remote Sensing Data data |
| CN108645342A (en) * | 2018-04-25 | 2018-10-12 | 国交空间信息技术（北京）有限公司 | A kind of road width extracting method based on path locus and high resolution image |
| CN108921076A (en) * | 2018-09-21 | 2018-11-30 | 南京信息工程大学 | Road face crack disease self-adaption constant false-alarm detection method based on image |
| CN109785307A (en) * | 2019-01-09 | 2019-05-21 | 武汉大学 | A kind of unmanned plane image road Damage assessment method based on vector guidance |
| CN109828999A (en) * | 2019-01-24 | 2019-05-31 | 厦门大学 | Road service system detection recognition method after calamity based on the cross-domain city big data of multi-source |
| CN110176000A (en) * | 2019-06-03 | 2019-08-27 | 斑马网络技术有限公司 | Road quality detection method and device, storage medium, electronic equipment |
| CN111260637A (en) * | 2020-01-19 | 2020-06-09 | 石家庄铁道大学 | Asphalt road transverse fracture detection method based on regional difference |
| CN111582280A (en) * | 2020-05-11 | 2020-08-25 | 吉林省森祥科技有限公司 | Deep data fusion image segmentation method for multispectral rescue robot |
| CN112241806A (en) * | 2020-07-31 | 2021-01-19 | 深圳市综合交通运行指挥中心 | Road damage probability prediction method, device terminal equipment and readable storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110169686A1 (en) * | 2008-05-29 | 2011-07-14 | Gian Luca Eusebi Borzelli | Target detection in a sar-imaged sea area |
| CN102831423A (en) * | 2012-07-26 | 2012-12-19 | 武汉大学 | SAR (synthetic aperture radar) image road extracting method |
-
2013
- 2013-12-30 CN CN201310750917.9A patent/CN103714339B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110169686A1 (en) * | 2008-05-29 | 2011-07-14 | Gian Luca Eusebi Borzelli | Target detection in a sar-imaged sea area |
| CN102831423A (en) * | 2012-07-26 | 2012-12-19 | 武汉大学 | SAR (synthetic aperture radar) image road extracting method |
Non-Patent Citations (4)
| Title |
|---|
| C. HAN ET AL.: "Road Extraction from High-Resolution SAR Image on Urban Area", 《2006 IEEE INTERNATIONAL SYMPOSIUM ON GEOSCIENCE AND REMOTE SENSING》 * |
| 徐阳 等: "基于道路网矢量数据的遥感影像道路损毁检测", 《测绘通报》 * |
| 梅珍 等: "SAR图像道路检测及其仿真", 《计算机仿真》 * |
| 陈世荣 等: "基于高分辨率遥感影像的汶川地震道路损毁评估", 《遥感学报》 * |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104050473B (en) * | 2014-05-20 | 2019-07-19 | 中国人民解放军理工大学 | A kind of road data extracting method based on rectangular neighborhood analysis |
| CN104050473A (en) * | 2014-05-20 | 2014-09-17 | 中国人民解放军理工大学 | Road data extraction method based on rectangular neighborhood analysis |
| CN104091369B (en) * | 2014-07-23 | 2017-02-22 | 武汉大学 | Unmanned aerial vehicle remote-sensing image building three-dimensional damage detection method |
| CN104091369A (en) * | 2014-07-23 | 2014-10-08 | 武汉大学 | Unmanned aerial vehicle remote-sensing image building three-dimensional damage detection method |
| CN104268879A (en) * | 2014-09-28 | 2015-01-07 | 民政部国家减灾中心 | Physical building quantity damage evaluation method based on remote sensing multi-spectral images |
| CN104268879B (en) * | 2014-09-28 | 2017-03-01 | 民政部国家减灾中心 | Building quantity of goods produced Damage assessment method based on remote sensing multispectral image |
| CN104408437A (en) * | 2014-12-08 | 2015-03-11 | 电子科技大学 | Road detecting method based on synthetic aperture radar |
| CN104408437B (en) * | 2014-12-08 | 2018-09-21 | 电子科技大学 | A kind of Approach for road detection based on synthetic aperture radar |
| CN106169086A (en) * | 2016-07-21 | 2016-11-30 | 武汉大学 | High-resolution optical image damage method for extracting roads under navigation data auxiliary |
| CN106169086B (en) * | 2016-07-21 | 2019-04-09 | 武汉大学 | High-resolution optical image under navigation data auxiliary damages method for extracting roads |
| CN106295699A (en) * | 2016-08-11 | 2017-01-04 | 北京师范大学 | A kind of earthquake Damage assessment method and apparatus based on high-definition remote sensing data |
| CN108229273A (en) * | 2017-02-27 | 2018-06-29 | 北京市商汤科技开发有限公司 | Multilayer neural network model training, the method and apparatus of roadway characteristic identification |
| CN108388828A (en) * | 2017-07-13 | 2018-08-10 | 中国科学院遥感与数字地球研究所 | A kind of seashore wetland land cover pattern information extracting method of comprehensive multi- source Remote Sensing Data data |
| CN107886505A (en) * | 2017-11-08 | 2018-04-06 | 电子科技大学 | A kind of synthetic aperture radar airfield detection method based on line primitives polymerization |
| CN108645342A (en) * | 2018-04-25 | 2018-10-12 | 国交空间信息技术（北京）有限公司 | A kind of road width extracting method based on path locus and high resolution image |
| CN108921076A (en) * | 2018-09-21 | 2018-11-30 | 南京信息工程大学 | Road face crack disease self-adaption constant false-alarm detection method based on image |
| CN108921076B (en) * | 2018-09-21 | 2022-02-18 | 南京信息工程大学 | Pavement crack disease self-adaptive constant false alarm detection method based on image |
| CN109785307A (en) * | 2019-01-09 | 2019-05-21 | 武汉大学 | A kind of unmanned plane image road Damage assessment method based on vector guidance |
| CN109828999A (en) * | 2019-01-24 | 2019-05-31 | 厦门大学 | Road service system detection recognition method after calamity based on the cross-domain city big data of multi-source |
| CN110176000A (en) * | 2019-06-03 | 2019-08-27 | 斑马网络技术有限公司 | Road quality detection method and device, storage medium, electronic equipment |
| CN111260637A (en) * | 2020-01-19 | 2020-06-09 | 石家庄铁道大学 | Asphalt road transverse fracture detection method based on regional difference |
| CN111582280A (en) * | 2020-05-11 | 2020-08-25 | 吉林省森祥科技有限公司 | Deep data fusion image segmentation method for multispectral rescue robot |
| CN111582280B (en) * | 2020-05-11 | 2023-10-17 | 吉林省森祥科技有限公司 | Data deep fusion image segmentation method for multispectral rescue robot |
| CN112241806A (en) * | 2020-07-31 | 2021-01-19 | 深圳市综合交通运行指挥中心 | Road damage probability prediction method, device terminal equipment and readable storage medium |
| CN112241806B (en) * | 2020-07-31 | 2021-06-22 | 深圳市综合交通运行指挥中心 | Road damage probability prediction method, device terminal equipment and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103714339B (en) | 2017-01-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103714339B (en) | SAR image road damaging information extracting method based on vector data | |
| Chamberlain et al. | A route-based visibility analysis for landscape management | |
| Pradhan et al. | A GIS-based back-propagation neural network model and its cross-application and validation for landslide susceptibility analyses | |
| Castillo et al. | The normalized topographic method: an automated procedure for gully mapping using GIS | |
| Rastiveis et al. | Automated extraction of lane markings from mobile LiDAR point clouds based on fuzzy inference | |
| Bowles et al. | Discovering marine terraces using airborne LiDAR along the Mendocino-Sonoma coast, northern California | |
| KR102399089B1 (en) | A system and method for flood damage detection using satellite image information | |
| KR20200091331A (en) | Learning method and learning device for object detector based on cnn, adaptable to customers' requirements such as key performance index, using target object merging network and target region estimating network, and testing method and testing device using the same to be used for multi-camera or surround view monitoring | |
| CN111626170B (en) | Image recognition method for railway side slope falling stone intrusion detection | |
| JP5500388B2 (en) | Shooting position specifying system, shooting position specifying program, and shooting position specifying method | |
| CN106022341A (en) | High resolution optical remote sensing image post-disaster water body information extracting method and system | |
| WO2023124442A1 (en) | Method and device for measuring depth of accumulated water | |
| Mora et al. | Small landslide susceptibility and hazard assessment based on airborne lidar data | |
| Rogers et al. | VEdge_Detector: automated coastal vegetation edge detection using a convolutional neural network | |
| Weidner et al. | Classifying rock slope materials in photogrammetric point clouds using robust color and geometric features | |
| Zhao et al. | Combining ICESat-2 photons and Google Earth Satellite images for building height extraction | |
| CN114494870A (en) | Double-time-phase remote sensing image change detection method, model construction method and device | |
| CN116246169A (en) | SAH-Unet-based high-resolution remote sensing image impervious surface extraction method | |
| Che et al. | Efficient segment-based ground filtering and adaptive road detection from mobile light detection and ranging (LiDAR) data | |
| Gerke et al. | Graph-supported verification of road databases | |
| Khaddari et al. | A comparative analysis of analytical hierarchy process and fuzzy logic modeling in flood susceptibility mapping in the Assaka Watershed, Morocco | |
| CN103927523B (en) | Fog level detection method based on longitudinal gray features | |
| Zhang et al. | Building footprint and height information extraction from airborne LiDAR and aerial imagery | |
| CN116205522B (en) | Landslide vulnerability evaluation method and system based on multidimensional CNN coupling | |
| Ren et al. | Geological disaster detection from remote sensing image based on experts' knowledge and image features |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |