CN105447875A - Automatic geometric correction method for electronic topographical map - Google Patents

Abstract

The present invention belongs to the technical field of development and application of geographic information technologies, and particularly relates to an automatic geometric correction method for an electronic topographical map. The method comprises the following steps: step 1. opening a scanned electronic topographical map in a grid format, naming the scanned electronic topographical map as a new map sheet number, carrying out calculation to obtain four-corner map border points, and generating four-corner latitude and longitude coordinates; step 2. performing template matching, performing preliminary matching on external map border four corners, and carrying out calculation by a similarity calculation method, so as to obtain map border point map plane coordinates; step 3. by the above steps, already obtaining four pairs of GCP points, transferring a path of the topographical map and the GCP point pairs into an ENVI (The Environment for Visualizing Images) system to perform registration, and completing automatic geometric correction of the electronic topographical map. Compared with the conventional manual operation method, the method is high in efficiency and precision and can greatly reduce manual labor; and the method can meet requirements of users for middle-scale topographical maps, and has a well-defined actual application prospect.

Description

A kind of electronic topographic map automatic geometric correction method

Technical field

The invention belongs to exploitation and the operation technique field of geographical information technology, particularly relate to a kind of electronic topographic map automatic geometric correction method.

Background technology

Electronic topographic map, as the fundamental geological data commonly used in a kind of research and design effort, carries out geometry correction after needing to change into the form of electronic image, could apply in a computer.GCP (Groundcontrolpoint, ground control point) point required for geometry correction is normally made up of the image coordinate of 4 neatline intersection points (i.e. Marginal point) and latitude and longitude coordinates.The latitude and longitude coordinates of Marginal point is given value, and corresponding image coordinate is general by manually measuring at present, and not only automaticity is low, and precision is low, the requirement of inapplicable contemporary work.At present, realize electronic topographic map autoregistration, first the unique point of known coordinate is to locate, and corner Marginal point (central point of crosshair) is known, key is the central point allowing computing machine automatically find these four crosshairs in certain sequence, then obtain image coordinate, form GCP point with known geographic coordinate, complete the coupling of topomap.If but the crosshair point of crossing of electronic topographic map is directly looked for by the method for pattern-recognition, no matter obtain straight-line intersection by HOUGH scheduling algorithm, or search for strong angle point by Harris scheduling algorithm, also or use the algorithm of images match, qualified result can considerably beyond four.If not there is deformation during topomap scanning, then likely determined the position of corner Marginal point by geometric relationship, but scanning map generation distortion is absolute.Therefore, in fact Marginal point be not easy to identify with obtain.

Summary of the invention

Goal of the invention of the present invention: utilize computer picture and vision technique and geographical information technology, proposes the method for a kind of secondary template coupling to realize the Automatic Calibration Technique of electronic topographic map; And in conjunction with opencv2.4.9, Envi4.8, under visualstudio2010 development environment, use C Plus Plus programmed method, solve the difficulty existed in background technology.The reliable reference mark of final acquisition, completes the automatic geometric correction of topomap.

Technical scheme of the present invention: a kind of electronic topographic map automatic geometric correction method, comprises the steps:

Step one, open the scanning electron topomap of a grid format, with number name of new map sheet, if the filename obtained does not meet the Naming conventions of new map sheet number, will miscue be jumped out; If the filename obtained meets the Naming conventions of new map sheet number, calculate according to this new map sheet number and obtain corner Marginal point, generated the latitude and longitude coordinates of corner right-angled intersection point by this map sheet number;

Step 2, template matches: the CV_TM_SQDIFF method utilizing template matches, outer border corner is mated for the first time, pass through crosshair point of crossing and the Marginal point of angle point and known longitude and latitude in corner after coupling, relatively-stationary distance is calculated by similarity calculating method and obtains Marginal point figure areal coordinate; In secondary template coupling, a square ROI region of interest centered by the drawing coordinate points obtained for the first time is set; In this region, use crosshair template again to mate, obtain Marginal point figure areal coordinate; Step 3, in ENVI+IDL environment, complete registration: through above step, four pairs of GCP points are obtained, by the path of topomap and GCP point to importing ENVI system into, utilize ENVI to have function envi_register_doit by oneself and carry out registration, ENVI_PROJ_CREATE function and establishment latitude and longitude coordinates system, finally complete electronic topographic map automatic geometric correction.

The scanning electron topomap of the grid format described in step one is .Jpg or .bmp or .GIFf formatted file.Described in step one, the standard of new map sheet Naming conventions foundation is: in Dec, 1992, and China has promulgated " national fundamental GIS framing and numbering GB/T13989-92 " file, and in March, 1993 comes into effect.

In step one, known new map sheet number calculates the longitude and latitude of this map sheet southwest Marginal point, is calculated as follows the longitude and latitude of this map sheet southwest Marginal point:

B＝(b-31)*6°+(d-1)*△B；

L＝(a-1)*4°+(4°/△L-c)*△L；

In formula: the longitude of certain point in B-------map sheet;

The latitude of certain point in L-------map sheet;

Topographic maps framing through difference required by △ B-------;

The meridional difference of topographic maps framing required by △ L-------;

Numerical code corresponding to the character code of latitude zone, a-------1:100 ten thousand topographic sheet place;

The numerical code of b-------1:100 ten thousand topographic sheet place longitude zone;

The line number of topographic maps required by c-------after 1:100 ten thousand topomap figure number;

The row number of topographic maps required by d-------after 1:100 ten thousand topomap figure number.

The preferred difference of two squares method of similarity calculating method in step 2.

Beneficial effect of the present invention: the present invention proposes a kind of new topomap auto-correction method, its core is secondary template matching algorithm, determine corner Marginal point image coordinate by this method, in conjunction with the map sheet number Marginal point theoretical coordinate calculated, form GCP point right, complete the automatic geometric correction of electronic topographic map, with the Measures compare of traditional-handwork operation, efficiency is high, and precision is also high, greatly can alleviate hand labor; For the demand that can meet user medium scale topomap, there is clear and definite actual application prospect.

Accompanying drawing explanation

Below in conjunction with accompanying drawing, the present invention is further illustrated.Accompanying drawing 1 is topomap automatic calibration process flow diagram of the present invention; Accompanying drawing 2 is topography image template: wherein a: crosshair; B: upper right outer border angle; C: outer border angle, bottom right; D: outer border angle, upper left; E: outer border angle, lower-left; Fig. 3 opens in Gloabalmapper to correct rear topomap and Stochastic choice milimeter grid point of crossing, and theoretical coordinate and true coordinate compare.

Embodiment

Embodiment 1, electronic topographic map automatic geometric correction method

One, materials and methods

1, material: the research object that this research uses is electronic topographic map, the material of use is the medium scale scanning electron topomap of 1: 5 within the scope of Xinjiang ten thousand and 1: 10 ten thousand.

2, method: the method for (1) secondary template coupling and thinking

Secondary template coupling use image matching algorithm, concrete grammar is, first mates the corner of outer border, outer border corner be feature obviously and be unique, each coupling all almost can reach the accuracy rate of 100%.Determine mapborder corner, the position of each outer border frame angle angle point and known nearest Marginal point (crosshair point of crossing) is fixing, be easy to like this determine a ROI (regionofinterested, region of interest), in this ROI region, re-use crosshair template matches, because only have the Marginal point of the known longitude and latitude of a right-angled intersection point-namely in this region, matching accuracy rate also can reach the accuracy rate of 100%.

(2) realization of concrete electronic topographic map automatic correction system

Development environment: operating system is 32-bit Windows 7 professional version, visual programming tools selects visualstudio2010, video and image development kit select opencv2.4.9, development language mainly uses C++, simultaneously after generating GCP (ground control point) file, use IDL (interactive data language, InteractiveDataLanguage) programming language in Envi4.8, call the geometry correction that own function completes electronic topographic map.System flow as shown in Figure 1.

Concrete steps one, map sheet number generate the latitude and longitude coordinates of corner right-angled intersection point

Open the scanning electron topomap of a grid format, under the environment of visualstudio2010 and C++, intercept filename, filename is with number name of new map sheet, if the filename obtained does not meet the Naming conventions of new map sheet number, to miscue be jumped out, if name is correct, then calculate according to new map sheet number the latitude and longitude coordinates obtaining corner Marginal point (being a crosshair intersection point).

Topomap identifies with figure number and the map title, and people find required drawing by searching concordance list, and map sheet number is in order to the management of topomap, preservation and occurring.In Dec, 1992, China has promulgated " national fundamental GIS framing and numbering GB/T13989-92 " file, in March, 1993 comes into effect, and specifies the standard of new map sheet number.According to this standard, known map sheet number calculates the longitude and latitude of this map sheet southwest Marginal point, is calculated as follows the longitude and latitude of this map sheet southwest Marginal point:

B＝(b-31)*6°+(d-1)*△B；

L＝(a-1)*4°+(4°/△L-c)*△L；

In formula: the longitude of certain point in B-------map sheet;

The latitude of certain point in L-------map sheet;

Topographic maps framing through difference required by △ B-------;

The meridional difference of topographic maps framing required by △ L-------;

Numerical code corresponding to the character code of latitude zone, a-------1:100 ten thousand topographic sheet place;

The numerical code of b-------1:100 ten thousand topographic sheet place longitude zone;

The line number of topographic maps required by c-------after 1:100 ten thousand topomap figure number;

The row number of topographic maps required by d-------after 1:100 ten thousand topomap figure number;

Step 2, template matches

Opens is a widely used computer vision storehouse of increasing income, and it provides a lot of function, achieves a lot of computer vision algorithms make, and algorithm all has from the most basic image and Video processing to senior recognition of face and relates to.In image procossing (IMGPROCMODULE) module, template matches (TemplateMatching) function can be realized, in Opens, images match can do the images match of man-to-man images match and one-to-many, because in topomap, outer border frame angle is unique, so adopt man-to-man image matching method here.

Similarity calculating method conventional in template matches has:

1) Squaredifferencematchingmethod (method=CV_TM_SQDIFF) difference of two squares method

$R(x,y)=\underset{x^{\′},y^{\′}}{\Σ}{\left(T\right(x^{\′},y^{\′})-I(x+x^{\′},y+y^{\′}\left)\right)}^2$

2) Correlationmatchingmethods (method=CV_TM_CCORR) Image Matching

$R(x,y)=\underset{x^{\′},y^{\′}}{\Σ}\left(T\right(x^{\′},y^{\′})\·I(x+x^{\′},y+y^{\′}\left)\right)$

3) Correlationcoefficientmatchingmethods (method=CV_TM_CCOEFF) related coefficient

$R(x,y)=\underset{x^{\′},y^{\′}}{\Σ}\left(T^{\′}\right(x^{\′},y^{\′})\·I(x+x^{\′},y+y^{\′}\left)\right)$

4) normalized form (object is the impact in order to reduce illumination) of above-mentioned three kinds of methods

5) SumofAbsolutedifferenceSAD absolute difference

6) the match is successful counts for MatchPixelCountMPC.By to the contrast of several method application result, find that the effect of difference of two squares method is best.

First utilize the CV_TM_SQDIFF method of template matches, outer border corner is mated, after coupling, obtain Marginal point image coordinate by the relatively-stationary distance calculating in crosshair point of crossing (Marginal point) of angle point in corner and known longitude and latitude.Because scintigram has certain distortion, calculate acquisition coordinate to overlap with Marginal point, but this is first coupling just, in secondary template coupling, arrange a square ROI region of interest centered by the image coordinate obtained for the first time point, length and width are that 30-40 pixel is advisable (300ppi image).In this region, use crosshair template again to mate, just can obtain the Marginal point image coordinate that precision is very high.

Step 3, in ENVI+IDL environment, complete registration

Through above step, four pairs of GCP points are obtained, because registration needs to carry out in ENVI+IDL environment, so need the path of topomap and GCP point importing ENVI system into.In visualstudio, conventional C# and C++ carrys out the secondary development that Callable IDL carries out ENVI, and this research uses the function of OPENCV to do template matches, and therefore programming language employs C++.

Under visual c++, the mode of Callable IDL has following several mode: 1. Callable technology; 2. IDLDrawWidget assembly: 3. COM_IDL_CONNECT assembly; 4. object exports assistant.The IDLDrawWidget assembly method selected and conveniently import parameter into carries out the contact between VC++ and ENVI+IDL, create the Widget_Draw assembly using IDL, comprise the method for Callable IDL function, supported data transmits, thus can be implemented in program interface and show the figure of IDL and the function of image.

IDLDrawWidget assembly possesses following three features: 1. directly can show the direct method figure under IDL or object method image; 2. the response of mouse and KeyEvent is provided; 3. data or variable transferring can be carried out with other language.Parameter utilizes ENVI to have function envi_register_doit by oneself to carry out registration after importing into, ENVI_PROJ_CREATE function and establishment latitude and longitude coordinates system, use ENVI_CONVERT_PROJECTION_CONRDINATES function to be converted into BJ54 Coordinate System system when needing.Be converted to GEOTIFF form from ENVi internal grid form .IMG and then need data conversion function ENVI_OUTPUT_TO_EXTERNAL_FORMAT.

Embodiment 2, from 12, Habahe County, xinjiang topomap: the template of Stochastic choice outer border corner and a crosshair template 12-45-5-first, 12-45-5-second, 12-45-5-third, 12-45-5-fourth, 12-45-6-first, 12-45-6-second, 12-45-6-third, 12-45-6-fourth, 12-45-17-first, 12-45-17-second, 12-45-17-third, 12-45-17-fourth, as shown in Figure 2, a map 12-45-17-second is have selected again from 12 topomap, new map sheet number is L45E003010, and image pixel is 300ppi.

The dialog box of MFC creates, as shown in Figure 3, interface is very succinct, as long as open image, after showing, determine to correct, wait for the image can seeing geometry correction for a moment under outgoing route, here have selected the GEOTIFF formatted file exporting BJ54 Coordinate System system, in Globalmap, open the effect having a look at correction.

The milimeter grid point coordinate theoretical value of table 1 Stochastic choice and comparing of actual value.

From the comparing of the milimeter grid point coordinate theoretical value of Stochastic choice and actual value, can find out, no matter be horizontal ordinate or ordinate, actual value after correction and theoretical coordinate absolute error are all within 10 meters, this is for the topomap of 1:5 ten thousand or 1:10 ten thousand, and precision is all the needs that can meet real work.

Claims (5)

1. an electronic topographic map automatic geometric correction method, comprises the steps:

Step one, open the scanning electron topomap of a grid format, with number name of new map sheet, if the filename obtained does not meet the Naming conventions of new map sheet number, will miscue be jumped out; If the filename obtained meets the Naming conventions of new map sheet number, calculate according to this new map sheet number and obtain corner Marginal point, generated the latitude and longitude coordinates of corner right-angled intersection point by this map sheet number;

Step 2, template matches: the CV_TM_SQDIFF method utilizing template matches, outer border corner is mated for the first time, pass through crosshair point of crossing and the Marginal point of angle point and known longitude and latitude in corner after coupling, relatively-stationary distance is calculated by similarity calculating method and obtains Marginal point figure areal coordinate; In secondary template coupling, a square ROI region of interest centered by the drawing coordinate points obtained for the first time is set; In this region, use crosshair template again to mate, obtain Marginal point figure areal coordinate;

Step 3, in ENVI+IDL environment, complete registration: through above step, four pairs of GCP points are obtained, by the path of topomap and GCP point to importing ENVI system into, utilize ENVI to have function envi_register_doit by oneself and carry out registration, ENVI_PROJ_CREATE function and establishment latitude and longitude coordinates system, finally complete electronic topographic map automatic geometric correction.

2. the method for claim 1, the scanning electron topomap of the grid format described in step one is .Jpg or .bmp or .GIFf formatted file.

3. the method for claim 1, described in step one, the standard of new map sheet Naming conventions foundation is: in Dec, 1992, and China has promulgated " national fundamental GIS framing and numbering GB/T13989-92 " file, and in March, 1993 comes into effect.

4. the method for claim 1, in step one, known new map sheet number calculates the longitude and latitude of this map sheet southwest Marginal point, is calculated as follows the longitude and latitude of this map sheet southwest Marginal point:

B=（b-31）*6°+（d-1）*△B；

L=（a-1）*4°+（4°/△L-c）*△L；

In formula: the longitude of certain point in B-------map sheet;

The latitude of certain point in L-------map sheet;

Topographic maps framing through difference required by △ B-------;

The meridional difference of topographic maps framing required by △ L-------;

Numerical code corresponding to the character code of latitude zone, a-------1:100 ten thousand topographic sheet place;

The numerical code of b-------1:100 ten thousand topographic sheet place longitude zone;

The line number of topographic maps required by c-------after 1:100 ten thousand topomap figure number;

The row number of topographic maps required by d-------after 1:100 ten thousand topomap figure number.

5. the method for claim 1, the preferred difference of two squares method of similarity calculating method in step 2.