CN103426153A - Unmanned aerial vehicle remote sensing image quick splicing method - Google Patents

Unmanned aerial vehicle remote sensing image quick splicing method Download PDF

Info

Publication number
CN103426153A
CN103426153A CN201310314306XA CN201310314306A CN103426153A CN 103426153 A CN103426153 A CN 103426153A CN 201310314306X A CN201310314306X A CN 201310314306XA CN 201310314306 A CN201310314306 A CN 201310314306A CN 103426153 A CN103426153 A CN 103426153A
Authority
CN
China
Prior art keywords
level
grid
aerial vehicle
remote sensing
unmanned aerial
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
Application number
CN201310314306XA
Other languages
Chinese (zh)
Other versions
CN103426153B (en
Inventor
李勇
张南峰
杨敬锋
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangzhou Zhong Ke Yun map Intelligent Technology Co., Ltd.
Original Assignee
Guangzhou Institute of Geography of GDAS
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guangzhou Institute of Geography of GDAS filed Critical Guangzhou Institute of Geography of GDAS
Priority to CN201310314306.XA priority Critical patent/CN103426153B/en
Publication of CN103426153A publication Critical patent/CN103426153A/en
Application granted granted Critical
Publication of CN103426153B publication Critical patent/CN103426153B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Image Processing (AREA)

Abstract

The invention discloses an unmanned aerial vehicle remote sensing image quick splicing method and relates to the technical field of image processing. The method includes the following steps: step 1, performing low-altitude aerial photographing by an imaging system mounted on an unmanned aerial vehicle; step 2, performing multilevel grid partitioning on a mass of sequence images; step 3, splicing and mapping corresponding sequence images in each level in a level-by-level manner; step 4, processing at a later stage. A CUDA (compute unified device architecture) algorithm is introduced, so that image splicing speed can be greatly increased; a CA (cellular automaton) model is utilized to optimize an SIFT (scale invariant feature transform) feature point algorithm, so that image matching efficiency and accuracy are improved.

Description

The quick joining method of a kind of unmanned aerial vehicle remote sensing images
Technical field
The present invention relates to image processing technique, especially a kind of quick joining method of unmanned aerial vehicle remote sensing images.
Background technology
With respect to satellite remote sensing and manned airborne remote sensing, the unmanned plane low-altitude remote sensing more fast, flexibly, especially for fields such as calamity emergencies, has satellite remote sensing and the incomparable advantage of manned airborne remote sensing.The unmanned plane low-altitude remote sensing can especially be provided for earthquake relief work by the high-definition remote sensing information the most timely that provides of secondary disaster (such as the checked-up lake monitoring) etc., and provides Information Assurance and Data support for the decision-making of earthquake relief work.Because unmanned aerial vehicle remote sensing images boat sheet quantity is many, inclination angle is large and irregular, the factor such as the longitudinal overlap degree is irregular, the coordinate position error is large, make that unmanned plane Image Matching difficulty is large, speed is slow, precision is low, have influence on the series of problems of image subsequent treatment.Unmanned plane low-altitude remote sensing data for magnanimity, traditional experience disposal route can not meet on a large scale, extraction and the analysis of high efficiency quick mapping and disaster information, and this disaster detection just, the disaster relief and the active demand of rebuilding in disaster-hit areas institute, so just need to work out the splicing of a kind of unmanned aerial vehicle remote sensing images accurately and efficiently, processing and drawing methods.
At present, aspect the research of unmanned aerial vehicle remote sensing images splicing processing method, mostly concentrate on the precision aspect that how to improve Image Matching, as the matching process based on feature etc.Unmanned plane image splicing method based on feature generally has following step: after (1) converts the coloured image read in to gray level image, then carry out feature point extraction on gray level image, and go out the yardstick of each unique point according to the metric space image calculation; (2) unique point oneself had, adopt certain computing method, obtains the principal direction of each unique point, makes algorithm have rotational invariance; (3), in order to allow unique point distinguish and come mutually, need, according to neighborhood territory pixel and specific describing method, generate descriptor to each unique point; (4) according to the descriptor of unique point, mated, using descriptor, very approaching point is to as same place; (5) owing to there will be unavoidably the mistake coupling according to descriptors match, need to be to overdue rejecting.In the unmanned plane image processing, algorithm with the most use is exactly SIFT (Scale Invariant Feature Transform) algorithm at present.General by introducing the SIFT algorithm, and add some constraint conditions to be mated.But people find in actual applications, although SIFT algorithm and combinational algorithm thereof can be obtained good matching effect, before image plane in enormous quantities, because self algorithm complex is high, memory consumption is large, and the deficiency such as very long consuming time is difficult to practicality.For these problems now existing researcher proposed some and improved one's methods, propose to adopt the SIFT algorithm of simplifying to carry out the analysis of image overlap degree, the SIFT algorithm of the employing also had based on GPU, the processing time of further accelerating the SIFT algorithm.But, the most basic or complexity and efficiency that need to change algorithm itself.
Process the main following problem that exists for large batch of unmanned plane sequential images at present: (1) is in image processing in enormous quantities, processing speed is slow, and memory consumption is very large, especially Bian SIFT algorithm, consuming time quite long, also there are a lot of improved places in algorithm itself, as met the demand of the quick mappings such as calamity emergency, is further improved; (2) in the splicing of unmanned plane sequential images, can produce cumulative errors, cause the image spliced later that serious distortion can occur, especially at parallax on larger image, how reducing stitching error, make splicing effect more desirable, is current problem in the urgent need to address; (3) Image Matching unique point skewness hooks, and a large amount of match points are assembled in the place of texture-rich, and the place that texture is sparse does not have match point, causes the relative orientation precision low, the image joint weak effect; (4) for the characteristics of image itself, do not carry out matching algorithm design etc.
Summary of the invention
Unmanned plane sequential images splicing for above-mentioned magnanimity, the deficiency that matching process based on feature exists that processing speed is slow, memory consumption is large, cumulative errors is many, splicing effect is poor etc., the invention provides the quick joining method of a kind of unmanned aerial vehicle remote sensing images, be intended to realize that unmanned plane sequential images to magnanimity splices fast, processing with become figure, and improve the precision of image joint.
For realizing above purpose, the technical scheme that the present invention takes is:
The quick joining method of a kind of unmanned aerial vehicle remote sensing images comprises the following steps:
Step 1, unmanned plane are carried imaging system and are carried out low latitude and take photo by plane, to obtain a plurality of boat sheets with magnanimity sequential images;
Step 2, described magnanimity sequential images is carried out to multi-layer gridding subregion based on the tile pyramid model, described multi-layer is arranged in order from bottom to the top, and every level at least comprises a grid;
Step 3, successively level is spliced rear one-tenth figure to a magnanimity sequential images corresponding in every level; Wherein, the described method that magnanimity sequential images corresponding in every level is spliced is:
The binding character of setting up the magnanimity sequential images in each grid CA(Cellular Automaton cellular automaton that automatically develops) model; And
The conversion of employing SIFT(Scale-invariant feature transform yardstick invariant features) the unique point algorithm is mated the magnanimity sequential images; Simultaneously
Magnanimity sequential images in grid in every level is adopted to CUDA(Compute Unified Device Architecture universal computing device framework) algorithm processes simultaneously;
Step 4, described one-tenth figure is carried out to post-processed.
Number of grid in each level reduces from bottom to the top successively.
Described grid is rectangle, and the length of side of the grid in every level equates.
The grid length of side in the last layer level is the integral multiple of the grid length of side in next level, and wherein, last layer level and next level are adjacent two-layer in described multi-layer, and the last layer level is that bottom or last layer level are than the close bottom of next level.
Each grid at least comprises a boat sheet, and the grid number in the last layer level equals the quantity of next level Air China sheet.
Described in step 3 successively the level for to carry out successively from the bottom to the top layer, and the magnanimity sequential images after the last layer level is processed enters in the grid of next level according to its coordinate, wherein, last layer level and next level are adjacent two-layer in described multi-layer, and the last layer level is that bottom or last layer level are than the close bottom of next level.
Post-processed described in step 4 at least comprises the registration of one-tenth figure, correction and toning.
Also comprise between step 1 and step 2 described boat sheet is carried out to geometric correction and coordinate registration.
When unmanned plane is carried out and to be taken photo by plane task, its longitudinal overlap degree > 30%, sidelapping degree 15%-30%, strip deformation degree<3%, swing angle<6 degree.
The present invention compared with prior art, has following advantage:
1, the present invention introduces the CUDA algorithm, it is a kind of new GPGPU (General Purpose computing on Graphics Processing Units general-purpose computations graphic process unit) technology, suppose to have N grid, comprise M in each grid and open the boat sheet, the number of threads of simultaneously being calculated is N * M * 8, so adopt CUDA to carry out the coupling of unmanned aerial vehicle remote sensing images in enormous quantities, can greatly improve the speed that data are processed.Through test, when boat sheet quantity is 1000 (take photo by plane area approximately 50 square kilometres), with classic method relatively, can raising efficiency approximately 2 times by this method; When boat sheet quantity more than 5000 or the area of taking photo by plane while surpassing 200 square kilometres, can raising efficiency approximately 4 times by this method.Concrete parameter index can be according to the difference of computer hardware configuration and different.The present invention is significant in the application in the fields such as calamity emergency for unmanned aerial vehicle remote sensing, the quick mapping of unmanned plane aerial images data can provide first hand data supporting for Disaster relief countermeasure, and disaster detection, the disaster relief and rebuilding in disaster-hit areas are had to important help.
2,, for each grid of each level, utilize restrictive evolution CA model automatically and in conjunction with SIFT unique point algorithm, image mated.Cellular automaton is all discrete power systems of a time and space, each cellular (Cell) be dispersed in regular grid is got limited discrete state, follow same effect rule, do synchronous the renewal according to definite local rule, a large amount of cellulars form the evolution of dynamic system by simple the interaction.Utilize the CA model to be optimized SIFT unique point algorithm, can further improve efficiency and the precision of Image Matching.By setting up the CA model of image in each grid, i.e. every 8 images interactions that image spatially is adjacent simultaneously, thus form dynamic matching relationship.
3, the present invention introduces the tile pyramid model, this model is a kind of multi-level model, from the bottom to the top layer, number of grid is fewer and feweri, but the geographic area scope meaned is constant, thereby avoid adopting simple concatenation former thereby cause the very large problem of spliced image positioning precision error because of cumulative errors.
The accompanying drawing explanation
The process flow diagram that Fig. 1 is the quick joining method of unmanned aerial vehicle remote sensing images of the present invention;
The structural drawing that Fig. 2 is multi-layer gridding subregion embodiment, wherein, empty frame representative boat sheet or spliced image, real frame represents grid.
Embodiment
Below in conjunction with the drawings and specific embodiments, content of the present invention is described in further details.
Embodiment
Please refer to shown in Fig. 1, a kind of unmanned aerial vehicle remote sensing images is the method for splicing fast, and it comprises the following steps:
S10, unmanned plane carry imaging system to carry out low latitude and takes photo by plane, to obtain a plurality of boat sheets with magnanimity sequential images.
Before carrying out step S20, must carry out to the original boat sheet of taking photo by plane processing early stage (operations such as geometric correction and coordinate registration) to improve efficiency and the precision of magnanimity sequential images coupling.
S20, the magnanimity sequential images of taking photo by plane is carried out to multi-layer gridding subregion based on the tile pyramid model, this multi-layer is arranged in order from bottom to the top, and every level at least comprises a grid.
For the picture of taking photo by plane in large zone, because data volume is very many, not only efficiency is low for simple joining method, and, due to the cumulative errors reason, can cause spliced image positioning precision error very large.The present invention introduces the tile pyramid model, and this model is a kind of multi-level model, and from the bottom to the top layer, number of grid is fewer and feweri, but the geographic area scope meaned is constant.Concrete grammar is as follows:, by being divided into one by one little square net, there are multiple boat sheets in the zone of taking photo by plane in each grid; Grid adopts square structure, and the length of side of the grid that every level is interior equates, but the boat sheet quantity comprised with the grid in layer possibility different (harmonies while processing in order to splice, preferably interior grid comprises identical boat sheet with layer); According to the regional size of taking photo by plane, determine and be divided into several levels, by bottom, to top layer, number of grid reduces step by step.The grid size of supposing the first level (bottom) is a * a (a is the ground floor grid square length of side), the grid size of the second level is b * b (b is the second layer grid square length of side), ... .., the grid size of n level (top layer) is that (N is the n layer grid square length of side to N * N, n >=2), pass between a, b and N is a<b<N, and b is that integral multiple, the N of a is the integral multiple of a and b.The total quantity that the quantity of bottom level Air China sheet is the unmanned plane boat sheet of taking photo by plane, the grid number of bottom equals the sum of second layer Air China sheet, by that analogy.Concrete gridding partition method as shown in Figure 2, in Fig. 2, the first level (bottom) has 16 grids, each grid comprises 16 boat sheets, has 4 grids in the second level, and each grid comprises 4 boat sheets, the 3rd level (being top layer here) has 1 grid, this grid comprises 4 boat sheets, the spliced map of top side in the Fig. 2 finally formed (figure), and concrete joining method refers to shown in step S30.
S30, successively level is spliced and one-tenth figure a magnanimity sequential images corresponding in every level.Successively level, for to carry out successively from the bottom to the top layer, specifically comprises:
S31, magnanimity sequential images corresponding in the first level (bottom) is spliced.The method of splicing is:
S311, the binding character of setting up the magnanimity sequential images in each grid CA model that automatically develops.Cellular automaton is all discrete power systems of a time and space, each cellular (Cell) be dispersed in regular grid is got limited discrete state, follow same effect rule, do synchronous the renewal according to definite local rule, a large amount of cellulars form the evolution of dynamic system by simple the interaction.
S312, employing SIFT unique point algorithm are mated the magnanimity sequential images.SIFT unique point algorithm is current algorithm with the most use in the unmanned plane image processing, no longer with regard to its processing procedure, is repeated here.Carry out step S313 with step S312 simultaneously.
S313, to the magnanimity sequential images in all grids in every level, adopt the CUDA algorithm to process simultaneously.
Integrating step S313 and step S312 are known, in the present invention, based on the CUDA algorithm, all grids that comprise in each level are carried out to the coupling that SIFT unique point algorithm carries out sequential images simultaneously.Can realize supposing to have N grid to carry out matching treatment between a plurality of grids, multiple images simultaneously, comprise M in each grid and open image, the number of threads of simultaneously being calculated is N * M * 8.Adopt CUDA to carry out the coupling of unmanned aerial vehicle remote sensing images in enormous quantities, can greatly improve the speed that data are processed.
Integrating step S311 and step S312 are known, by setting up the CA model of image in each grid, i.e. and every 8 images interactions that image spatially is adjacent simultaneously, thus form dynamic matching relationship.Utilize the CA model to be optimized SIFT unique point algorithm, can further improve efficiency and the precision of Image Matching.
The image that the first level has been spliced rear formation enters according to its coordinate in the grid of the second level, then performs step S32.
Step S32, the grid in the second level is spliced, the method for splicing is identical with step S311-S313, and the image that the second level has been spliced rear formation enters according to its coordinate in the grid of the 3rd level, by that analogy, until carry out step S33.
Step S33, the image of the grid of last level (top layer) is spliced, joining method is identical with step S32 with step S31.
After image joint in step S34, last level grid completes, the achievement of formation is last one-tenth figure.
Step S40, last one-tenth figure is carried out to a series of post-processed could use, post-processed mainly comprises that the achievement formed after post-processed is only final available image to the registration of one-tenth figure, correction, toning etc.
The method that the concrete test of take is spliced the present invention fast as example describes.
Should concrete test adopt the fixed-wing unmanned plane to carry digital camera and taken photo by plane in Shunde District, Fushan City, Guangdong Province, 806 square kilometres of the regional total areas of taking photo by plane.The unmanned plane technical parameter is as follows: captain 1788mm, span 2260mm, the high 430mm of machine, draw before power arrangement, the power configuration twin-tub is opposed, control module UP20 flight control system, 2 kilograms of equipment loads, 15 kilograms of take-off weights, 120 kilometers/hour of maximal raties, 90 kilometers/hour of cruise speed, 3500 meters of maximum ceilings, 100 minutes cruising time.Take photo by plane and adopt Canon EOS 400D digital camera, 1,100 ten thousand pixels, 24 millimeters tight shots with digital camera.Longitudinal overlap degree 30% when task is taken photo by plane in execution, sidelapping degree 30%, strip deformation degree<3%, swing angle<6 degree.This is taken photo by plane and sets 600 meters of flying heights, 0.144 meter of ground resolution, and monolithic is area on the spot: 652*435=0.28362 square kilometre, this 11520 of common acquisition boat sheet of taking photo by plane.
At first apply professional image software original boat sheet is carried out to geometric correction and coordinate registration.Then multi-layer gridding subregion is carried out in the whole zone of taking photo by plane, be divided into three grades: the first level boat sheet quantity is 11520, setting boat sheet quantity in each grid of the first level is 36, totally 320 grids, and to enter the image quantity of second layer level after having mated be 320 to the first level; Setting boat sheet quantity in each grid of the second level is 16, totally 20 grids, and to enter the image quantity of the 3rd level after having mated be 20 to the second level; Image of formation after the 3rd 20 of levels coupling completes.SIFT unique point algorithm after the restrictive evolution CA model optimization automatically of utilization is mated the boat sheet the order by the first level, the second level, the 3rd level, finally forms a striograph.The image that last coupling is formed is manually adjusted, coordinate registration, color processing etc., finally forms end result.
Above-listed detailed description is for the illustrating of possible embodiments of the present invention, and this embodiment is not in order to limit the scope of the invention, and the equivalence that all the present invention of disengaging do is implemented or change, all should be contained in the protection domain of this case.

Claims (10)

1. the quick joining method of unmanned aerial vehicle remote sensing images, is characterized in that, comprises the following steps:
Step 1, unmanned plane are carried imaging system and are carried out low latitude and take photo by plane, to obtain a plurality of boat sheets with magnanimity sequential images;
Step 2, described magnanimity sequential images is carried out to multi-layer gridding subregion based on the tile pyramid model, described multi-layer is arranged in order from bottom to the top, and every level at least comprises a grid;
Step 3, successively level is spliced rear one-tenth figure to a magnanimity sequential images corresponding in every level; Wherein, the described method that magnanimity sequential images corresponding in every level is spliced is:
The binding character of setting up the magnanimity sequential images in each grid CA model that automatically develops; And
Adopt SIFT unique point algorithm to be mated the magnanimity sequential images; Simultaneously
To magnanimity sequential images in the grid in every level, adopt the CUDA algorithm to process simultaneously;
Step 4, described one-tenth figure is carried out to post-processed.
2. the quick joining method of unmanned aerial vehicle remote sensing images according to claim 1, is characterized in that, the number of grid in each level reduces from bottom to the top successively.
3. the quick joining method of unmanned aerial vehicle remote sensing images according to claim 1 and 2, is characterized in that, described grid is rectangle, and the length of side of the grid in every level equates.
4. the quick joining method of unmanned aerial vehicle remote sensing images according to claim 3, it is characterized in that, the grid length of side in the last layer level is the integral multiple of the grid length of side in next level, wherein, last layer level and next level are adjacent two-layer in described multi-layer, and the last layer level is that bottom or last layer level are than the close bottom of next level.
5. the quick joining method of unmanned aerial vehicle remote sensing images according to claim 4, is characterized in that, each grid at least comprises a boat sheet, and the grid number in the last layer level equals the quantity of next level Air China sheet.
6. the quick joining method of unmanned aerial vehicle remote sensing images according to claim 4, it is characterized in that, described in step 3, successively level for to carry out successively from the bottom to the top layer, and the magnanimity sequential images after last layer level processing enters in the grid of next level according to its coordinate.
7. the quick joining method of unmanned aerial vehicle remote sensing images according to claim 1 and 2, it is characterized in that, described in step 3 successively the level for to carry out successively from the bottom to the top layer, and the magnanimity sequential images after the last layer level is processed enters in the grid of next level according to its coordinate, wherein, last layer level and next level are adjacent two-layer in described multi-layer, and the last layer level is that bottom or last layer level are than the close bottom of next level.
8. the quick joining method of unmanned aerial vehicle remote sensing images according to claim 1, is characterized in that, post-processed described in step 4 at least comprises the registration of described one-tenth figure, correction and toning.
9. the quick joining method of unmanned aerial vehicle remote sensing images according to claim 1, is characterized in that, also comprises between step 1 and step 2 described boat sheet is carried out to geometric correction and coordinate registration.
10. the quick joining method of unmanned aerial vehicle remote sensing images according to claim 1, is characterized in that, described unmanned plane is carried out while taking photo by plane task, its longitudinal overlap degree > 30%, sidelapping degree 15%-30%, strip deformation degree<3%, swing angle<6 degree.
CN201310314306.XA 2013-07-24 2013-07-24 The quick joining method of a kind of unmanned aerial vehicle remote sensing images Active CN103426153B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201310314306.XA CN103426153B (en) 2013-07-24 2013-07-24 The quick joining method of a kind of unmanned aerial vehicle remote sensing images

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201310314306.XA CN103426153B (en) 2013-07-24 2013-07-24 The quick joining method of a kind of unmanned aerial vehicle remote sensing images

Publications (2)

Publication Number Publication Date
CN103426153A true CN103426153A (en) 2013-12-04
CN103426153B CN103426153B (en) 2016-01-20

Family

ID=49650847

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201310314306.XA Active CN103426153B (en) 2013-07-24 2013-07-24 The quick joining method of a kind of unmanned aerial vehicle remote sensing images

Country Status (1)

Country Link
CN (1) CN103426153B (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107918938A (en) * 2017-11-16 2018-04-17 凌云光技术集团有限责任公司 A kind of matching process and device of point set and point set
CN110033411A (en) * 2019-04-12 2019-07-19 哈尔滨工业大学 The efficient joining method of highway construction scene panoramic picture based on unmanned plane
CN111080525A (en) * 2019-12-19 2020-04-28 成都海擎科技有限公司 Distributed image and primitive splicing method based on SIFT (Scale invariant feature transform) features
CN111768339A (en) * 2020-06-29 2020-10-13 广西翼界科技有限公司 Rapid splicing method for aerial images of unmanned aerial vehicle
CN111860205A (en) * 2020-06-29 2020-10-30 成都数之联科技有限公司 Forest fire evaluation method based on multi-source remote sensing image and grid and storage medium
CN112802082A (en) * 2021-04-13 2021-05-14 中国测绘科学研究院 Motion recovery structure method suitable for large-scale scene

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120039547A1 (en) * 2008-06-24 2012-02-16 Microsoft Corporation Variable resolution images
CN102663057A (en) * 2012-03-02 2012-09-12 苏州武大影像信息工程研究院有限责任公司 Method for managing multisource-isomerism aviation remote sensing data

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120039547A1 (en) * 2008-06-24 2012-02-16 Microsoft Corporation Variable resolution images
CN102663057A (en) * 2012-03-02 2012-09-12 苏州武大影像信息工程研究院有限责任公司 Method for managing multisource-isomerism aviation remote sensing data

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
李劲澎 等: "无人机影像SIFT特征匹配的集群并行处理方法", 《测绘科学技术学报》, vol. 29, no. 6, 31 December 2012 (2012-12-31), pages 440 - 444 *
袁修国 等: "基于GPU的变型SIFT算子实时图像配准", 《计算机科学》, vol. 38, no. 3, 31 March 2011 (2011-03-31), pages 300 - 303 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107918938A (en) * 2017-11-16 2018-04-17 凌云光技术集团有限责任公司 A kind of matching process and device of point set and point set
CN110033411A (en) * 2019-04-12 2019-07-19 哈尔滨工业大学 The efficient joining method of highway construction scene panoramic picture based on unmanned plane
CN110033411B (en) * 2019-04-12 2021-01-12 哈尔滨工业大学 High-efficiency road construction site panoramic image splicing method based on unmanned aerial vehicle
CN111080525A (en) * 2019-12-19 2020-04-28 成都海擎科技有限公司 Distributed image and primitive splicing method based on SIFT (Scale invariant feature transform) features
CN111768339A (en) * 2020-06-29 2020-10-13 广西翼界科技有限公司 Rapid splicing method for aerial images of unmanned aerial vehicle
CN111860205A (en) * 2020-06-29 2020-10-30 成都数之联科技有限公司 Forest fire evaluation method based on multi-source remote sensing image and grid and storage medium
CN111860205B (en) * 2020-06-29 2024-03-19 成都数之联科技股份有限公司 Forest fire evaluation method based on multisource remote sensing images and grids and storage medium
CN112802082A (en) * 2021-04-13 2021-05-14 中国测绘科学研究院 Motion recovery structure method suitable for large-scale scene
CN112802082B (en) * 2021-04-13 2021-07-20 中国测绘科学研究院 Motion recovery structure method suitable for large-scale scene

Also Published As

Publication number Publication date
CN103426153B (en) 2016-01-20

Similar Documents

Publication Publication Date Title
CN103426153B (en) The quick joining method of a kind of unmanned aerial vehicle remote sensing images
CN103226838A (en) Real-time spatial positioning method for mobile monitoring target in geographical scene
CN113286129B (en) Inspection method and system for photovoltaic power station
CN110189405B (en) Live-action three-dimensional modeling method considering building density
JP7065477B2 (en) Disaster situation judgment system and disaster judgment flight system
CN107545538A (en) A kind of Panorama Mosaic method and device based on unmanned plane
CN114332385A (en) Monocular camera target detection and spatial positioning method based on three-dimensional virtual geographic scene
CN103744661A (en) Multi-sensor data integration processing method and system of minimum altitude unmanned aerial vehicle
Qu et al. Construction site monitoring using uav oblique photogrammetry and bim technologies
CN112633661A (en) BIM-based emergency dispatching command method, system, computer equipment and readable medium
CN111178283A (en) Unmanned aerial vehicle image-based ground object identification and positioning method for established route
CN111696156B (en) Remote sensing image coordinate conversion method without control point
CN113741510A (en) Routing inspection path planning method and device and storage medium
CN116192044A (en) Fault photovoltaic panel numbering and positioning method and device, electronic equipment and storage medium
CN116223511A (en) Distributed roof photovoltaic module defect diagnosis method and device based on unmanned aerial vehicle automatic inspection
CN116129067A (en) Urban live-action three-dimensional modeling method based on multi-source geographic information coupling
EP3282351A1 (en) System and method for facilitating an inspection process
CN104180794A (en) Method for treating texture distortion area of digital orthoimage
CN103808309A (en) Three-dimensional aerial photograph forest measurement method for unmanned aerial vehicle
CN111982077B (en) Electronic map drawing method and system and electronic equipment
CN116129064A (en) Electronic map generation method, device, equipment and storage medium
CN114550016B (en) Unmanned aerial vehicle positioning method and system based on context information perception
Ge et al. Monitoring dynamic deformation of building using unmanned aerial vehicle
Han et al. Research on fine 3D modeling technology of tall buildings based on UAV Photogrammetry
Wang et al. Mountainous village relocation planning with 3D GIS virtual imaging space model and neural network

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
TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20191224

Address after: 510000 room 708, building 1, No. 20, taihegang Road, Yuexiu District, Guangzhou City, Guangdong Province

Patentee after: Guangzhou Zhong Ke Yun map Intelligent Technology Co., Ltd.

Address before: 510070, 100 martyrs Road, Guangdong, Guangzhou, one of the compound

Patentee before: Guangzhou Institute of Geography