Summary of the invention
In view of the foregoing, it is necessary to provide a kind of streetscape to set up system and streetscape method for building up, the seamless link of three-dimensional stereo model can be realized.
A kind of streetscape sets up system, this system includes: an image acquiring module, this image acquiring module is for obtaining the scene image absorbed in each place by least three camera head, in each place, this at least three camera head absorbs multiple scene image respectively and can synthesize the image of multiple 360 degree, and this scene image includes the range information in this place between each object and corresponding camera head;One object detecting module, this object detecting module is used for utilizing object detecting technology, obtains each point in scene image and arrive the distance of corresponding camera head in each place scene image obtained;One direction acquisition module, direction acquisition module is for obtaining the direction of those camera heads shooting of an electronic compass detecting;One position acquisition module, this position acquisition module is for obtaining the place of those camera heads picked-up image of a positioner detecting;And a model building module, this model building module each distance putting those camera heads in the scene image shot according to different location and different directions and those scene images sets up multiple three-dimensional stereo model, and determine two three-dimensional stereo models that the domain of the existence shot in adjacent place is overlapping, two places are carried out phase adduction and is averaged coordinate, according to this average coordinates, the same position point of these two three-dimensional stereo models is alignd and synthesizes, the three-dimensional stereo model shot other different locations and different directions successively synthesizes, set up a three-dimensional streetscape model.
A kind of streetscape method for building up, the method includes: obtain the scene image absorbed in each place by least three camera head, multiple scene images that this at least three camera head absorbs respectively in each place can synthesize multiple 360 degree of images, and this scene image includes the range information in this place between each object and corresponding camera head;Utilize object detecting technology, each place scene image obtained obtains each point in scene image and arrives the distance of corresponding camera head;Obtain the direction of the camera head shooting of an electronic compass detecting;Obtain the place of those camera heads picked-up image of a positioner detecting;In the scene image shot according to different location and different directions and those scene images, each point sets up multiple three-dimensional stereo model to the distance of those camera heads, and determine two three-dimensional stereo models that the domain of the existence shot in adjacent place is overlapping, two places carry out phase adduction be averaged coordinate and alignd by the same position point of these two three-dimensional stereo models according to this average coordinates, the scene image shot other different locations and different directions successively synthesizes, and sets up a three-dimensional streetscape model.
The present invention obtains the scene image of these three camera head picked-ups by this image acquiring module, this object detecting module obtains each point in scene image and arrives the distance of camera head, direction acquisition module obtains the direction of the camera head shooting of this electronic compass detecting, this position acquisition module obtains the place of those camera heads picked-up image of this positioner detecting, this model building module sets up three-dimensional stereo model according to those scene images and distance, determine two three-dimensional stereo models that domain of the existence that adjacent place shoots is overlapping, thus two three-dimensional stereo models are alignd and synthesize, successively other three-dimensional stereo model is synthesized.Thus realize the seamless link of three-dimensional stereo model.
Detailed description of the invention
As it is shown in figure 1, be the streetscape of an embodiment of the present invention application block diagram of setting up system 1.A three-dimensional streetscape image is set up in the place that those camera heads 2 that this streetscape sets up the orientation of the shooting of those camera heads 2 of scene image that system 1 absorbs, electronic compass 3 detecting according at least three camera head 2 and a positioner 4 determines absorb image.
This camera head 2 is used for absorbing scene, and generates scene image.In the present embodiment, described camera head 2 is a kind of degree of depth camera (Depth-sensingCamera), such as time flight (TimeofFlight, TOF) camera, for absorbing scene image, this scene image includes the depth of view information of subject.The depth of view information of described subject refers to the range information of subject each point and camera head.Owing to camera head 2 is when photographic subjects thing, the signal of certain wavelength will be launched, camera head 2 can be reflexed to when signal runs into object, time difference between launching according to signal and receive can calculate the range information on object between each point and camera head 2, the most described camera head 2 can get the range information in scene image between subject each point and camera head 2, range information between subject each point and camera head 2 in i.e. depth of view information, hereinafter referred to as scene image.Wherein, as in figure 2 it is shown, this at least three camera head 2 is arranged on a concentric circular, and mutual distance is equal, for absorbing certain angle scope, for example, at least image of 120 degree of scopes, thus the image of this at least three camera head 2 picked-up can synthesize one 360 degree of scene images.
Described streetscape is set up system 1 and is included a memory element 11 and several functional modules 12.Described memory element 11 is for storing the scene image of camera head 2 picked-up.
This functional module 12 is stored in memory element 11 with the form of computer processing, and is configured to the execution of one or more processing unit, to complete the present invention.Refer to Fig. 3, in the present embodiment, these several functional modules 12 include image acquiring module 121, object detecting module 122, direction acquisition module 123, position acquisition module 124 and a model building module 125.
This image acquiring module 121 is for obtaining the scene image in each place in the street absorbed in each place by least three camera head 2.
This object detecting module 122 utilizes object detecting technology, obtains each point in scene image and arrive the distance of corresponding camera head 2 in each place scene image obtained.In the present embodiment, described object detecting technology can use Robust real-time objects detection method (RobustReal-timeObjectDetectionMethod).
Direction acquisition module 123 is for obtaining the direction of the shooting of those camera heads 2 of this electronic compass 3 detecting.
Those camera heads 2 that this position acquisition module 124 determines for obtaining this positioner 4 absorb the place of image.In the present embodiment, those camera heads 2, apart from close, can be considered a location point.
This model building module 125 determines different location and the scene image of different directions shooting, each point in the scene image shot according to those different locations and different directions and those scene images sets up multiple three-dimensional stereo model to the distance of camera head 2, and according to this positional information and scene image, determine in adjacent place, such as A point and B point, two three-dimensional stereo models that captured domain of the existence is overlapping, the position coordinates phase adduction of A point and B point is taken average coordinates C, according to this average coordinates C, the same position point of these two three-dimensional stereo models is alignd and synthesizes, the three-dimensional stereo model shot other different locations and different directions successively synthesizes, generate a three-dimensional streetscape model, thus realize 360 degree of seamless links of three-dimensional streetscape model.
In the present embodiment, this functional module 12 also includes an image analysing computer module 126.This image analysing computer module 126 is used for analyzing in this scene image whether there is on-fixed object, such as the people of movement, animal, or vehicle etc..Specific as follows:
Camera head 2, at same place repeatedly filmed image, obtains multiple image.Wherein, in the present embodiment, this camera head 2 is arranged on a roof, can slowly move and switch spot for photography, in same spot for photography, with relatively short period of time interval shooting multiple image, obtains multiple images in same place.In other embodiments, this camera head 2, after the image having shot all places, can repeat again to shoot repeatedly, obtains multiple images in same place.
This image analysing computer module 126 absorbs image place according to shooting orientation and the camera head 2 of the camera head 2 obtained, the distance of each camera head 2 point in multiple images that same place shoots to camera head 2 is carried out mutual comparison, when one of them or several images exist several points to distance differences of camera head 2 relative to other images, this image analysing computer module 126 determines and there is on-fixed object in this scene image.
In the present embodiment, this image analysing computer module 126 also can absorb image place according to the shooting orientation of the camera head 2 obtained and camera head 2, each camera head 2 X in multiple images that same place shoots or Y-direction image are contrasted, when one of them or time several images exist X or Y-direction image difference relative to other images, this image analysing computer module 126 determines and there is on-fixed object in this scene image.
When there is on-fixed object in this scene image, this image analysing computer module 126 deletes the on-fixed object in this scene image.The scene image of on-fixed object is there is not and sets up three-dimensional stereo model in this model building module 125 according to this.Wherein, this scene image that there is not on-fixed object can be the scene image after deleting on-fixed object or be the scene image that the most there is not on-fixed object.
In the present embodiment, this functional module 12 also includes a model analysis module 127.This model analysis module 127 obtains the pixel that the pixel of each place value in this scene image that there is not on-fixed object is corresponding, calculate the pixel average of pixel, and corresponding for this pixel pixel average is affixed on the relevant position of this three-dimensional streetscape model, generate a three-dimensional streetscape image, then in this image, the streetscape of optional position has the color of correspondence, as at position D, the color of this streetscape is blue, and at position E, the color of this streetscape is red, thus gives people's sense of reality.
Refer to Fig. 4, for the flow chart of the streetscape method for building up of an embodiment of the present invention.This streetscape method for building up is applied to streetscape as shown in Figure 1 and sets up in system 1.
In step S301, this image acquiring module 121 obtains the scene image in each place of the streetscape absorbed in each place by least three camera head 2.
In step s 302, this object detecting module 122 obtains each point in scene image in each place scene image obtained and arrives the distance of corresponding camera head 2.
In step S303, direction acquisition module 123 obtains this electronic compass 3 and detects the direction of those camera heads 2 shooting.
In step s 304, this position acquisition module 124 obtains those camera heads 2 that this positioner 4 determines and absorbs the place of image.
In step S305, this image analysing computer module 126 analyzes whether there is on-fixed object in this scene image.When this scene image exists on-fixed object, perform step S306.When this scene image does not exist on-fixed object, perform step S307.
In step S306, this image analysing computer module 126 deletes the on-fixed object in this scene image.
In step S307, this model building module 125 determines different location and the scene image of different directions shooting, each point in the scene image shot according to those different locations and different directions and those scene images sets up multiple three-dimensional stereo model to the distance of camera head 2, and according to this positional information and three-dimensional stereo model, determine in different location, two three-dimensional stereo models that captured domain of the existence is overlapping, the position coordinates phase adduction of different location is taken an average coordinates, according to this average coordinates, the same position point of these two three-dimensional stereo models is alignd and synthesizes, the three-dimensional stereo model shot other different locations and different directions successively synthesizes, generate a three-dimensional streetscape model.
In step S308, this model analysis module 127 obtains the pixel of each pixel in this scene image that there is not on-fixed object, calculate the corresponding pixel average of each pixel, and this corresponding pixel average of each pixel is affixed on the relevant position of this three-dimensional streetscape model, generate a three-dimensional streetscape image.
Finally it should be noted that, above example is only in order to illustrate technical scheme and unrestricted, although the present invention being described in detail with reference to preferred embodiment, it will be understood by those within the art that, technical scheme can be modified or equivalent, without deviating from the spirit and scope of technical solution of the present invention.