CN102855609A - Shallow underwater topography construction method integrating hyper-spectral data and sparse sonar data - Google Patents
Shallow underwater topography construction method integrating hyper-spectral data and sparse sonar data Download PDFInfo
- Publication number
- CN102855609A CN102855609A CN2012102681709A CN201210268170A CN102855609A CN 102855609 A CN102855609 A CN 102855609A CN 2012102681709 A CN2012102681709 A CN 2012102681709A CN 201210268170 A CN201210268170 A CN 201210268170A CN 102855609 A CN102855609 A CN 102855609A
- Authority
- CN
- China
- Prior art keywords
- data
- sonar data
- pixel
- remote sensing
- sparse
- 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
- 238000012876 topography Methods 0.000 title claims abstract description 28
- 238000010276 construction Methods 0.000 title claims abstract description 18
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 70
- 238000000034 method Methods 0.000 claims abstract description 27
- 238000012937 correction Methods 0.000 claims abstract description 7
- 230000009467 reduction Effects 0.000 claims description 38
- 238000001228 spectrum Methods 0.000 claims description 23
- 238000005259 measurement Methods 0.000 claims description 20
- 230000003595 spectral effect Effects 0.000 claims description 14
- 238000009792 diffusion process Methods 0.000 claims description 11
- 230000005855 radiation Effects 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 4
- 238000013144 data compression Methods 0.000 claims description 3
- 238000007621 cluster analysis Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 abstract description 8
- 230000011218 segmentation Effects 0.000 description 5
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000013179 statistical model Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 239000004215 Carbon black (E152) Substances 0.000 description 1
- 238000009360 aquaculture Methods 0.000 description 1
- 244000144974 aquaculture Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000013213 extrapolation Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 229930195733 hydrocarbon Natural products 0.000 description 1
- 150000002430 hydrocarbons Chemical class 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000012856 packing Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 239000002304 perfume Substances 0.000 description 1
- VYMDGNCVAMGZFE-UHFFFAOYSA-N phenylbutazonum Chemical compound O=C1C(CCCC)C(=O)N(C=2C=CC=CC=2)N1C1=CC=CC=C1 VYMDGNCVAMGZFE-UHFFFAOYSA-N 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000004064 recycling Methods 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Landscapes
- Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
Abstract
The invention relates to a shallow underwater topography construction method integrating hyper-spectral data and sparse sonar data and belongs to the technical field of underwater topography reconnaissance. According to the shallow underwater topography construction method, dimensions of a hyper-spectral remote sensing image are reduced by aid of a clustering center of sonar data, area division is performed on a low dimensional remote sensing image after dimension reducing, and interpolation is performed on sonar data inside each of the areas to obtain the underwater topography. According to the shallow underwater topography construction method, the hyper-spectral remote sensing image and sparse sonar data are organically combined, in the whole process, and two kinds of data compensate to solve the problem of underwater topography construction well. The remote sensing image and the sonar data are provided with coordinate information after geometric correction, and a certain fuzzy corresponding relation exists between a grayscale of the remote sensing image and the water depth, so that water depth in each depth homogeneous area changes slightly in area-divided remote sensing image, and interpolation results of sonar data are authentic.
Description
Technical field
The present invention relates to a kind of shallow water underwater landform construction method, especially relate to the underwater topography construction method of a kind of integrated high-spectral data and sparse sonar data, belong to underwater topography prospecting technique field.
Background technology
Frequent interactional area, the extra large land of the regional conduct of coastal, Offshore, the geological location is superior, all kinds of aboundresources, mankind's activity is violent.From the resources development and utilization angle, Offshore is the strategic Development Base of hydrocarbon resources, food resource, shipping resource, with this end in view carry out the job facilities such as ground coast protection, port and harbour construction, mudflat aquaculture, energy development, waterway development, cable laying all the utmost point need the submarine topography geomorphologic map of various accurate, different scales.
At present, obtaining of underwater topography data mainly comprises boat-carrying measurement and remote-sensing inversion.Wherein traditional boat-carrying acoustic measurement takes sounding device and the combined mode of positioning equipment to carry out the measurement of underwater topography usually, and sounding device is mainly sonar.Sonar can be divided into three kinds of modes by its principle of work: single-point type, single beam and multi-beam.Multi-beam sonar is measured and has been realized that the underwater topography face covers mapping, but the multi-beam sonar equipment price is too expensive, so the underwater topography data acquisition still uses single beam, single-point type sonar in a large number.Use single beam, when single-point type equipment carries out bathymetric surveying, carry out the depth measurement flight course planning according to the depth measurement feature on water, form the depth measurement band net line of going forward side by side and measure, the modes such as recycling interpolation extrapolation are carried out data encryption, thereby obtain test block underwater topography product.Because the equipment acquired results such as sonic echo depth finder are section survey, can't represent submarine topography, geomorphology information comprehensively, the in addition restriction of the conditions such as environment, time and funds, measurement data overlay area, the renewal frequency can not satisfy the requirement of the large regional normalization of oceanographic survey, real time implementation.
20 world end of the sixties, U.S.'s remote sensing sounding technology that begun one's study.Different spectrum, spatial resolution remotely-sensed data are on the increase in recent years, and shallow water depth of water inverse model is enriched constantly, can be divided into the forms such as theoretical explanation model, semiempirical model and statistical model according to the principle of its inverting institute foundation.
The theoretical explanation model take visible light in water body radiation transfer equation as the basis, by reasonable assumption abbreviation model, be aided with the inner multiple light mathematic(al) parameter of the water body of surveying gained and carry out depth of water inverting, the method has clear and definite physical significance, but water body internal optics difficult parameters can not widespread use with acquisition.
For the complicated and difficult difficult point that obtains of theoretical explanation model parameter, semiempirical model take light in water the attenuation principle as the basis, adopt the model such as water body attenuation and analyze the empirical parameter that returns gained to combine and realize the remote-sensing inversion of the depth of water, can be divided into single band model and multiband model according to the remote sensing wave band number that adopts.
No matter be take visible light in water body radiation transfer equation as the theoretical model on basis, also be based on the semiempirical model of the attenuation principle of light in water body, because the water body optical characteristics is very complicated, the optical parametric observed reading is difficult to obtain, affect the many factors of depth of water inversion accuracy, so the model inversion precision is not high.
Statistical model is widely used in the depth of water inverting with its simple inverse model, higher inversion accuracy, but the method needs a certain amount of actual measurement water depth value, thereby has restricted to a certain extent the popularization of model.
Summary of the invention
The present invention wants the technical solution problem to be: the shallow water underwater landform construction method of a kind of integrated high-spectral data and sparse sonar data is provided, in conjunction with high-spectral data, carries out accurately the structure of shallow water underwater landform with more sparse sonar data.
In order to solve above technical matters, the invention provides the shallow water underwater landform construction method of a kind of integrated high-spectral data and sparse sonar data, may further comprise the steps:
Step 1, data preparation---through the conversion of absolute radiation value, the rejecting of clutter wave band, image rotation, bad line reparation and striped removal, the contrary rotation of image, geometry correction and the disposal of gentle filter target in hyperspectral remotely sensed image, with the sparse sonar data of coordinate information and Water Depth Information, described sparse sonar data is corresponding with the target in hyperspectral remotely sensed image coordinate range;
Step 2, seek sonar data the cluster centre point---the water depth value to sonar data carries out cluster, and obtain all kinds of central values, the approaching measurement point of water depth value and described central value is as all kinds of central points;
Step 3, the semi-supervised dimensionality reduction of high spectrum--process in the remote sensing image that obtains through step 1, the pixel identical with described center point coordinate is as the sample point of semi-supervised dimensionality reduction, utilize semi-supervised Laplce's dimension-reduction algorithm of considering space length to carry out dimensionality reduction to processing the remote sensing image that obtains through step 1, obtain the low-dimensional remote sensing image;
The division of step 4, depth of water homogenous area---use the multi-scale division algorithm to cut apart to the low-dimensional remote sensing image, the low-dimensional remote sensing image is divided into some depth of water homogenous area;
Step 5, to the sonar data interpolation---at the inner information diffusion method that uses of each depth of water homogenous area sonar data is carried out interpolation, obtains underwater topography.
The present invention organically combines target in hyperspectral remotely sensed image and sparse sonar data, and in whole process, two kinds of data have solved the problem that underwater topography makes up well as a kind of complementation.At first the cluster centre point by sonar data carries out dimensionality reduction to target in hyperspectral remotely sensed image, then the low-dimensional remote sensing image behind the dimensionality reduction is carried out the zone and divides, and at each intra-zone sonar data is carried out interpolation at last and obtains underwater topography.Remote sensing image and sonar data through geometry correction all possess coordinate information, and there are certain fuzzy corresponding relation in the gray scale of remote sensing image and the depth of water, therefore can think behind the partition in the remote sensing image, in each depth of water homogenous area depth of water conversion less, the sonar data interpolation result is truer.The present invention has utilized this rule just cleverly, with two kinds of data the combine high precision that realized underwater topography, the low-cost structure.
For the sonar data that multi-beam face type is measured, the data low price that the present invention is used.With respect to traditional remote-sensing inversion model, the present invention need not to obtain water body optical parametric observed reading, and the factor that affects underwater topography structure precision is less, and the underwater topography precision of structure is higher.
Further, the present invention improves Laplce's dimension reduction method.Tradition Laplce dimension reduction method is considered spectrum intervals, ignored the spatial character between the remote sensing image picture element, and the present invention improves the associating distance that spectrum intervals is spectrum intervals and space length, and is specific as follows:
The semi-supervised dimensionality reduction of high spectrum of step 3 is for considering Laplce's dimensionality reduction of locus, and the distance threshold of dimension-reduction algorithm adopts associating Gauss distance, and associating Gauss range formula is as follows:
D wherein
IjThe associating distance that represents j pixel of i pixel to the,
Represent j pixel spectrum intervals of i pixel to the,
The space length that represents j pixel of i pixel to the,
In the formula,
The water depth value of i pixel of expression spectral information,
The water depth value of j pixel of expression spectral information,
The water depth value of i pixel of expression spectral information K-band,
The water depth value of j pixel of expression spectral information K-band, parameter σ
SThe control coefrficient of expression spectrum intervals;
The coordinate figure of i pixel of representation space information,
The coordinate figure of j pixel of representation space information,
The coordinate figure of i pixel of representation space information K-band,
J cell coordinate value of representation space information K-band, σ
LThe control coefrficient of representation space distance.
Consider that space length is about tens times of spectrum intervals, in the improved dimension-reduction algorithm of the present invention, the control coefrficient σ of spectrum intervals
SSpan be 1-2, the control coefrficient σ of space length
LSpan be 85-105.
Further, in the described step 2, by the k-means algorithm sonar data is carried out cluster analysis, seek the optimum cluster pattern, obtain all kinds of central values.
Further, in the described step 2, water depth value and described central value deviation ± 0.2m with interior measurement point as all kinds of central points.
For computation reduction, in the step 1 of the present invention, also to the following pre-service of described sparse sonar data process, sonar data is divided into regular shipping lines measurement point and pass point two parts, the regular shipping lines measurement point is carried out vacuate according to the picture dot size, use average as picture dot central point water depth value, again two parts data are merged the sparse sonar data that obtains after the data compression.
Innovative point of the present invention is: (1) for traditional non-linearity manifold dimension-reduction algorithm take vector space model as the basis, being applied to the remotely-sensed data acquired results distributes concentrated in the low dimensional manifold space, easily cause mixed minute phenomenon of classification, this method is introduced semi-supervised thought and the spatial relationship that embeds based on figure, a kind of semi-supervised Laplce's dimension-reduction algorithm of considering space length has been proposed, by regulating spectral space associating Gauss distance parameter, controlled sample point low dimensional manifold embeds, and then utilize semi-supervised Laplce's algorithm to obtain take the dimensionality reduction result of space of matrices model as the basis, experiment shows the more traditional Laplce's dimension-reduction algorithm of this arithmetic result, separability in the low dimensional manifold embedded space of all categories increases, and is beneficial to follow-up cutting apart and classification.(2) in the process of semi-supervised dimensionality reduction, use the Kmeans method that sonar point data are carried out repeatedly cluster, obtain optimum cluster centre as the sampled point of semi-supervised dimensionality reduction, so that the result behind the dimensionality reduction and Water Depth Information have higher correlativity.(3) method with diffusion of information is applied in the interpolation of sparse sonar point data, the method is by selecting the diffusion of information that sparse sample is entrained of suitable Fuzzy information diffusion model in the monitoring space, and find the solution the corresponding predicted value of future position by the theory of fuzzy set, make up the step of search circle when having broken away from common interpolating method and choosing data, interpolation of data sparse to observation station and that be placed with specific rule has preferably effect.(4) propose the new approaches that a kind of underwater topography makes up, namely utilize sparse banded sonar data and face to cover high-spectral data, take the integrated strategy of multi-source data, learn from other's strong points to offset one's weaknesses, reach and improve the purpose that underwater topography makes up precision.
Description of drawings
The present invention is further illustrated below in conjunction with accompanying drawing.
Fig. 1 is the target in hyperspectral remotely sensed image data of the embodiment of the invention.
Fig. 2 is the sonar data of the embodiment of the invention.
Fig. 3 is the image picture element among Fig. 1 corresponding to the sonar data cluster centre point that obtains of Fig. 2.
Fig. 4 is the as a result figure behind the semi-supervised dimensionality reduction of Fig. 1.
Fig. 5 is for carrying out the as a result figure after the multi-scale division to Fig. 4.
Fig. 6 utilizes information diffusion method that Fig. 2 is carried out the underwater topographic map that interpolation obtains.
Embodiment
The present embodiment makes up near the bay Tampa Bay underwater topography Gulfian, east coast of the united states, and the high-spectrum remote sensing data of employing is a scape in 2004, and these data are totally 242 wave bands, 30 meters of spatial resolutions, as shown in Figure 1; The sonar data that adopts is Tampa Bay single-point type sonar data in 2004, and at a distance of 500m, each sonar point is at a distance of 3.5m, as shown in Figure 2 on the course line between the adjacent course line of data.
The shallow water underwater landform construction method of the integrated high-spectral data of this example and sparse sonar data mainly may further comprise the steps:
Step 1, data are prepared.
1) high-spectral data is carried out pre-service.The used data of the present embodiment are through processes such as blotch removal, echo correction, background removal, radiation correcting, the recovery of bad pixel and picture quality detections.But in the practical application, still exist bad pixel need to carry out further pre-service, to reach the purpose that improves the image availability, concrete grammar is as follows:
1. absolute radiation value conversion.High-spectral data is convenient storage because the actual emanations value is less, multiply by certain coefficient on the original value basis, so need here all VNIR wave bands divided by all wave bands of 40, SWIR divided by 80.
2. wave band is rejected.Rejecting is without the wave band 1～7,58～7,255～242 of radiation calibration, overlapping wave band 56～57,77～78 and the wave band 121～127,167～178 that is subjected to the steam image.
3. bad line reparation and striped are removed.Vertical state carries out bad line reparation, striped is removed so that bad line and striped are in the image rotation.Method (the Goodenough DG that the people such as employing Goodenough propose, DykA, NiemannKO, et al Processing hyper-ion and ali for forest classification[J] .IEEE Transactionson Geoscience and Remote Sensing, 2003,41 (6): 1321-1331.) carry out automatic detection and the reparation of bad line.Adopt " overall situation is removed striped " method (" pre-service of EO-1Hyperion high-spectral data ", " sensor information " the 6th phase in 2005, author: TanBing Xiang, Li Zengyuan, Chen Erxue, Pang Yong) of the propositions such as the bright perfume of Tan to carry out the reparation of striped.Then image is carried out the contrary virgin state that rotates to be.
4. geometry correction.Basis is carried out geometric accurate correction through the TM remotely-sensed data with accurate geographic position information of registration to high-spectral data in this example.
5. smothing filtering.Be the stochastic error that the random variation of eliminating the image numeric data causes, adopt the 5x5 medium filtering to eliminate high frequency noise, obtain more level and smooth spatial spectral radiation numerical value, in order to rear depth of water inverting.
2) sonar data is carried out pre-service.
Packing density is 0.5m/point on the sonar course line, the highdensity measurement of course-and-bearing improves limited to inversion accuracy, but greatly aggravated the complicacy of calculating and the expense of internal memory, the extraction that the strategy of taking minute feature to resample carries out sonar data has reached the purpose that reduces computing time and space complexity.
Concrete grammar is as follows: sonar data is divided into regular shipping lines measurement point and pass point two parts, the regular shipping lines measurement point is carried out vacuate according to the picture dot size, use average as picture dot central point water depth value, again two parts data are merged the sparse sonar data that obtains after the data compression.
The cluster centre point of step 2, searching sonar data.Water depth value to sonar data carries out cluster, and obtains all kinds of central values, and the measurement point that water depth value and described central value approach is as all kinds of central points.
Seek the sonar data cluster centre as follows as the concrete grammar of semi-supervised sample point:
1) progressively increases cluster centre point number, and change initial random some number of cluster, ask repeatedly error mean in the cluster result class, choose the cluster centre number of AME minimum in the class as actual cluster centre number.
2) in step 1) under the given cluster centre, progressively increase initial random some number of cluster, until after the cluster in the accumulative total class of each bunch square error obviously increase;
3) select step 2) middle 2 points that add up error minimum in the class, and near these 2, constantly change cluster centre point number, with square error in the accumulative total class of approaching the possibility minimum, obtain optimum cluster centre point.
The process of cluster is as shown in table 1, in this example when the cluster centre number is 13, the 13rd the interior sample number of classification is 0, therefore give up 13, finally choosing the Optimal cluster centers number is 12,-11.1602 ,-9.3621 ,-3.8004 ,-13.9071 ,-15.1869 ,-8.2899 ,-19.7460 ,-24.5605 ,-5.1402 ,-10.7514 ,-12.1251 ,-6.7161 ,-16.4174 optimum initial random counting is 195, and corresponding cluster centre is as follows:.
Table 1
Step 3, the semi-supervised dimensionality reduction of high spectrum.Process the remote sensing image that obtains through step 1, stacked with the sonar data cluster centre point in the step 2, obtain the image picture element that intersects with cluster centre point, corresponding image picture element carries out semi-supervised dimensionality reduction take these pixel points as sample point as shown in Figure 3.
Tradition Laplce dimension reduction method is considered spectrum intervals, ignored the spatial character between the remote sensing image picture element, and the present embodiment improvement spectrum intervals is the associating distance of spectrum intervals and space length, utilize semi-supervised Laplce's dimension-reduction algorithm of considering space length to carry out dimensionality reduction to processing the remote sensing image that obtains through step 1, obtain the low-dimensional remote sensing image.
The semi-supervised dimensionality reduction of high spectrum of the present embodiment is for considering Laplce's dimensionality reduction of locus, and the distance threshold of dimension-reduction algorithm adopts associating Gauss distance, and associating Gauss range formula is as follows:
D wherein
IjThe associating distance that represents j pixel of i pixel to the,
Represent j pixel spectrum intervals of i pixel to the,
The space length that represents j pixel of i pixel to the,
In the formula,
The water depth value of i pixel of expression spectral information,
The water depth value of j pixel of expression spectral information,
The water depth value of i pixel of expression spectral information K-band,
The water depth value of j pixel of expression spectral information K-band, parameter σ
SThe control coefrficient of expression spectrum intervals;
The coordinate figure of i pixel of representation space information,
The coordinate figure of j pixel of representation space information,
The coordinate figure of i pixel of representation space information K-band,
J cell coordinate value of representation space information K-band, σ
LThe control coefrficient of representation space distance.
Distinguish dimensionality reduction for processing each piecemeal that conveniently whole high-spectral data is divided into the 100*100 size, the dimension that the present embodiment is set behind the dimensionality reduction is 2 dimensions, and the K neighborhood quantity of Laplce's dimension-reduction algorithm is set as 12.K field quantity is that (acquiescence is made as 8 or 12 for a parameter of Laplce's dimension-reduction algorithm, here choose 12), Laplce's dimensionality reduction basic thought is to describe a stream shape with a undirected authorized graph, guarantee the structure before behind dimensionality reduction each authorized graph can also keep original dimensionality reduction, because this authorized graph is a local syntople, so determine an adjacent region threshold, be the K adjacent region threshold.Concrete grammar is as follows:
1) reads in related data and comprise image data, spatial position data and sampling number certificate.
2) use non-supervisory Laplce's dimension-reduction algorithm to obtain sample point as semi-supervised sampled point.
3) consider that according to sample point the locus semi-supervised Laplce's dimension-reduction algorithm of use carries out dimensionality reduction to the high-spectral data after cutting apart, setting K neighborhood quantity is 12, and the spectrum intervals coefficient value is 1, and the space length coefficient value is 100.
1. at first traversal is treated the dimensionality reduction data set
Look for the Neighbor Points of each point, structure neighbour figure.
In, calculate each some X
iWith the spectrum between all the other points and the associating Gauss distance in space, seek with respect to each some X
i12 nearest points of Euclidean distance then be classified as the Neighbor Points of required point, if some X
iWith X
jAdjacency, then figure mid point X
iWith X
jBetween have the undirected limit of having the right.
2. use heat kernel function structure weight matrix W.In above-mentioned neighbour figure, for selecting a weights W in each bar limit
I, j, structure weight matrix W.If some X
iWith X
jBe adjacency, the weights of then establishing the limit are W
I, j=exp (|| X
i-Y
j||
2/ t), otherwise establish W
I, j=0.
3. according to sample point
The low-dimensional corresponding with sample point embeds
The Laplacian matrix is carried out piecemeal process, and bring original Laplacian dimensionality reduction equation into
In, ask for the low-dimensional embedded coordinate for the treatment of the dimensionality reduction data set
The dimensionality reduction result of each piecemeal is merged, and final merging view picture dimensionality reduction result out as shown in Figure 4.
The division of step 4, depth of water homogenous area.Use the multi-scale division algorithm to cut apart to the low-dimensional remote sensing image, the low-dimensional remote sensing image is divided into some depth of water homogenous area.
The present embodiment uses multi-scale segmentation method, and (the multi-scale segmentation method algorithm can be with reference to " a kind of OO multi-scale segmentation of remote sensing images method ", " Beijing Jiaotong University's journal: natural science edition " the 4th phase in 2007, author: Tan's thoroughfare continuous heavy rain, Liu Zhengjun, Shen Wei) result behind the dimensionality reduction is carried out multi-scale division, constantly adjust the yardstick threshold value of cutting apart, final by constantly adjusting the yardstick threshold value, finally choose 62 as the optimum segmentation yardstick, and carrying out on this basis the piecemeal merging, the segmentation result that obtains is as shown in Figure 5.
Step 5, to the sonar data interpolation.At the inner information diffusion method that uses of each depth of water homogenous area sonar data is carried out interpolation, obtain underwater topography, as shown in Figure 6.
Information diffusion method can be with reference to " information diffusion method and application thereof in the fuzzy message processing ", " Northwest University's journal (natural science edition) " 04 phase in 1992, author: Wang Jiading, Huang Chongfu.
At first use each piecemeal in the step 4 that the sonar data point is cut apart, sonar data in each piecemeal uses the method for diffusion of information to carry out interpolate value, then the result with each piecemeal diffusion of information is stitched together, and obtains the underwater topography information of whole example region.
In addition to the implementation, the present invention can also have other embodiments.All employings are equal to the technical scheme of replacement or equivalent transformation formation, all drop on the protection domain of requirement of the present invention.
Claims (6)
1. the shallow water underwater landform construction method of an integrated high-spectral data and sparse sonar data may further comprise the steps:
Step 1, data preparation---through the conversion of absolute radiation value, the rejecting of clutter wave band, image rotation, bad line reparation and striped removal, the contrary rotation of image, geometry correction and the disposal of gentle filter target in hyperspectral remotely sensed image, with the sparse sonar data of coordinate information and Water Depth Information, described sparse sonar data is corresponding with the target in hyperspectral remotely sensed image coordinate range;
Step 2, seek sonar data the cluster centre point---the water depth value to sonar data carries out cluster, and obtain all kinds of central values, the approaching measurement point of water depth value and described central value is as all kinds of central points;
Step 3, the semi-supervised dimensionality reduction of high spectrum---process in the remote sensing image that obtains through step 1, the pixel identical with described center point coordinate is as the sample point of semi-supervised dimensionality reduction, utilize semi-supervised Laplce's dimension-reduction algorithm of considering space length to carry out dimensionality reduction to processing the remote sensing image that obtains through step 1, obtain the low-dimensional remote sensing image;
The division of step 4, depth of water homogenous area---use the multi-scale division algorithm to cut apart to the low-dimensional remote sensing image, the low-dimensional remote sensing image is divided into some depth of water homogenous area;
Step 5, to the sonar data interpolation---at the inner information diffusion method that uses of each depth of water homogenous area sonar data is carried out interpolation, obtains underwater topography.
2. the shallow water underwater landform construction method of integrated high-spectral data according to claim 1 and sparse sonar data, it is characterized in that: the semi-supervised dimensionality reduction of high spectrum of described step 3 is for considering Laplce's dimensionality reduction of locus, the distance threshold of dimension-reduction algorithm adopts associating Gauss distance, and associating Gauss range formula is as follows:
D wherein
IjThe associating distance that represents j pixel of i pixel to the,
Represent j pixel spectrum intervals of i pixel to the,
The space length that represents j pixel of i pixel to the,
In the formula,
The water depth value of i pixel of expression spectral information,
The water depth value of j pixel of expression spectral information,
The water depth value of i pixel of expression spectral information K-band,
The water depth value of j pixel of expression spectral information K-band, parameter σ
SThe control coefrficient of expression spectrum intervals;
The coordinate figure of i pixel of representation space information,
The coordinate figure of j pixel of representation space information,
The coordinate figure of i pixel of representation space information K-band,
J cell coordinate value of representation space information K-band, σ
LThe control coefrficient of representation space distance.
3. the shallow water underwater landform construction method of integrated high-spectral data according to claim 2 and sparse sonar data, it is characterized in that: in the described step 2, by the k-means algorithm sonar data is carried out cluster analysis, seek the optimum cluster pattern, obtain all kinds of central values.
4. the shallow water underwater landform construction method of integrated high-spectral data according to claim 3 and sparse sonar data is characterized in that: in the described step 2, water depth value and described central value deviation ± 0.2m with interior measurement point as all kinds of central points.
5. the shallow water underwater landform construction method of integrated high-spectral data according to claim 4 and sparse sonar data, it is characterized in that: in the step 1, described sparse sonar data is through following pre-service, sonar data is divided into regular shipping lines measurement point and pass point two parts, the regular shipping lines measurement point is carried out vacuate according to the picture dot size, use average as picture dot central point water depth value, again two parts data are merged the sparse sonar data that obtains after the data compression.
6. the shallow water underwater landform construction method of integrated high-spectral data according to claim 5 and sparse sonar data is characterized in that: the control coefrficient σ of spectrum intervals
SSpan be 1-2, the control coefrficient σ of space length
LSpan be 85-105.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210268170.9A CN102855609B (en) | 2012-07-30 | 2012-07-30 | Shallow underwater topography construction method integrating hyper-spectral data and sparse sonar data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210268170.9A CN102855609B (en) | 2012-07-30 | 2012-07-30 | Shallow underwater topography construction method integrating hyper-spectral data and sparse sonar data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102855609A true CN102855609A (en) | 2013-01-02 |
| CN102855609B CN102855609B (en) | 2014-10-01 |
Family
ID=47402173
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210268170.9A Expired - Fee Related CN102855609B (en) | 2012-07-30 | 2012-07-30 | Shallow underwater topography construction method integrating hyper-spectral data and sparse sonar data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102855609B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103292784A (en) * | 2013-05-14 | 2013-09-11 | 哈尔滨工程大学 | Underwater surface topography visuality analyzing method based on sound ray track |
| WO2015003496A1 (en) * | 2013-07-08 | 2015-01-15 | 国家***第二海洋研究所 | Method for constructing submarine topography and geomorphology based on fusion of bathymetric data from multiple sources |
| CN109459759A (en) * | 2018-11-13 | 2019-03-12 | 中国科学院合肥物质科学研究院 | City Terrain three-dimensional rebuilding method based on quadrotor drone laser radar system |
| CN109544691A (en) * | 2018-11-05 | 2019-03-29 | 国家***第二海洋研究所 | Automatically the MF method of multi-source heterogeneous bathymetric data building high-resolution DBM is merged |
| CN111174912A (en) * | 2020-01-03 | 2020-05-19 | 南京大学 | Snapshot type dispersion ambiguity-resolving hyperspectral imaging method |
| CN113139337A (en) * | 2021-04-14 | 2021-07-20 | 扬州大学 | Partitioned interpolation processing method and device for lake terrain simulation |
| CN115235431A (en) * | 2022-05-19 | 2022-10-25 | 南京大学 | Shallow sea water depth inversion method and system based on spectrum layering |
| CN115564743A (en) * | 2022-10-18 | 2023-01-03 | 故宫博物院 | Method and device for extracting typical spectrum of paper and related equipment |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030076742A1 (en) * | 2001-08-28 | 2003-04-24 | Rowe Francis D. | Acoustic doppler channel flow measurement device |
| CN101034471A (en) * | 2007-03-29 | 2007-09-12 | 上海大学 | Landform transformation method for sonar remote sensing digital image of underwater digit land model |
-
2012
- 2012-07-30 CN CN201210268170.9A patent/CN102855609B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030076742A1 (en) * | 2001-08-28 | 2003-04-24 | Rowe Francis D. | Acoustic doppler channel flow measurement device |
| CN101034471A (en) * | 2007-03-29 | 2007-09-12 | 上海大学 | Landform transformation method for sonar remote sensing digital image of underwater digit land model |
Non-Patent Citations (2)
| Title |
|---|
| EDWARD H.BOSCH等: "Wavelet-based dimension reduction for hyperspectral image classification", 《ALGORITHMS AND TECHNOLOGIES FOR MULTISPECTRAL, HYPERSPECTRAL, AND ULTRASPECTRAL IMAGERY IX》 * |
| 辜智慧等: "气象观测站点稀疏地区的降水插值方法探讨", 《北京师范大学学报（自然科学版）》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103292784A (en) * | 2013-05-14 | 2013-09-11 | 哈尔滨工程大学 | Underwater surface topography visuality analyzing method based on sound ray track |
| WO2015003496A1 (en) * | 2013-07-08 | 2015-01-15 | 国家***第二海洋研究所 | Method for constructing submarine topography and geomorphology based on fusion of bathymetric data from multiple sources |
| US9361728B2 (en) | 2013-07-08 | 2016-06-07 | The Second Institute Of Oceanography, Soa | Submarine topography construction method based on multi-source water depth data integration |
| CN109544691B (en) * | 2018-11-05 | 2021-07-06 | 国家***第二海洋研究所 | MF (multi-frequency) method for automatically fusing multi-source heterogeneous water depth data to construct high-resolution DBM (database management system) |
| CN109544691A (en) * | 2018-11-05 | 2019-03-29 | 国家***第二海洋研究所 | Automatically the MF method of multi-source heterogeneous bathymetric data building high-resolution DBM is merged |
| CN109459759A (en) * | 2018-11-13 | 2019-03-12 | 中国科学院合肥物质科学研究院 | City Terrain three-dimensional rebuilding method based on quadrotor drone laser radar system |
| CN111174912A (en) * | 2020-01-03 | 2020-05-19 | 南京大学 | Snapshot type dispersion ambiguity-resolving hyperspectral imaging method |
| CN113139337A (en) * | 2021-04-14 | 2021-07-20 | 扬州大学 | Partitioned interpolation processing method and device for lake terrain simulation |
| CN113139337B (en) * | 2021-04-14 | 2023-09-22 | 扬州大学 | Partition interpolation processing method and device for lake topography simulation |
| CN115235431A (en) * | 2022-05-19 | 2022-10-25 | 南京大学 | Shallow sea water depth inversion method and system based on spectrum layering |
| CN115235431B (en) * | 2022-05-19 | 2024-05-14 | 南京大学 | Shallow sea water depth inversion method and system based on spectrum layering |
| CN115564743A (en) * | 2022-10-18 | 2023-01-03 | 故宫博物院 | Method and device for extracting typical spectrum of paper and related equipment |
| CN115564743B (en) * | 2022-10-18 | 2023-11-03 | 故宫博物院 | Method and device for extracting typical spectrum of paper and related equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102855609B (en) | 2014-10-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102855609B (en) | Shallow underwater topography construction method integrating hyper-spectral data and sparse sonar data | |
| Liu et al. | Patterns of the loop current system and regions of sea surface height variability in the eastern Gulf of Mexico revealed by the self‐organizing maps | |
| Wang et al. | Predictive mapping of soil total nitrogen at a regional scale: A comparison between geographically weighted regression and cokriging | |
| Kang et al. | Topographic mapping on large-scale tidal flats with an iterative approach on the waterline method | |
| Shi-Biao et al. | GIS-based and data-driven bivariate landslide-susceptibility mapping in the Three Gorges area, China | |
| Bhardwaj et al. | Generation of high-quality digital elevation models by assimilation of remote sensing-based DEMs | |
| CN107679476B (en) | Sea ice type remote sensing classification method | |
| CN104698462A (en) | Sea surface wind field fusion method for SAR (Synthetic Aperture Radar) based on variation | |
| Coleman et al. | Holes in the ocean: Filling voids in bathymetric lidar data | |
| CN116720310A (en) | Seawater depth inversion method integrating wave evolution numerical model | |
| Cheng et al. | Integration of hyperspectral imagery and sparse sonar data for shallow water bathymetry mapping | |
| Mohebzadeh et al. | Spatial downscaling of MODIS Chlorophyll-a with machine learning techniques over the west coast of the Yellow Sea in South Korea | |
| Yang et al. | Bathymetric inversion and mapping of two shallow lakes using Sentinel-2 imagery and bathymetry data in the central Tibetan Plateau | |
| An et al. | A review on the research progress of lake water volume estimation methods | |
| CN117523404B (en) | Method, device, terminal and storage medium for monitoring dynamic change of storage yard | |
| Migolet et al. | Estimation of the total dry aboveground biomass in the tropical forests of Congo Basin using optical, LiDAR, and radar data | |
| Utlu et al. | Comparison of morphometric characteristics of dolines delineated from TOPO-Maps and UAV-DEMs | |
| Marghany et al. | Simulation of shoreline change using AIRSAR and POLSAR C-band data | |
| Zhang et al. | A method to derive tidal flat topography in Nantong, China using MODIS data and tidal levels | |
| Zhao et al. | Water deep mapping from HJ-1B satellite data by a deep network model in the sea area of Pearl River Estuary, China | |
| Al-Mutairi et al. | Spatial enhancement of DEM using interpolation methods: a case study of Kuwait’s coastal zones | |
| Li | Dynamic monitoring algorithm of natural resources in scenic spots based on MODIS Remote Sensing technology | |
| Reddy | Remote sensing and GIS in digital terrain modeling | |
| Lewicka | Analysis of Terrain Modelling Methods in the Coastal Zone | |
| Baghdadi et al. | QGIS and Applications in Water and Risks |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20141001 Termination date: 20170730 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |