EP2201524A2 - Method for creating and/or updating textures of background object models, video monitoring system for carrying out the method, and computer program - Google Patents

Abstract

Video monitoring systems are used for camera-supported monitoring of relevant areas, and usually comprise a plurality of monitoring cameras placed in the relevant areas for recording monitoring scenes. The monitoring scenes may be, for example, parking lots, intersections, streets, plazas, but also regions within buildings, plants, hospitals, or the like. In order to simplify the analysis of the monitoring scenes by monitoring personnel, the invention proposes displaying at least the background of the monitoring scene on a monitor as a virtual reality in the form of a three-dimensional scene model using background object models. The invention proposes a method for creating and/or updating textures of background object models in the three-dimensional scene model, wherein a background image of the monitoring scene is formed from one or more camera images 1 of the monitoring scene, wherein the background image is projected onto the scene model, and wherein textures of the background object models are created and/or updated based on the projected background image.

Description

 Method for generating and / or updating textures of background object models, video surveillance systems for carrying out the

Procedure as well as computer program

State of the art

The invention relates to a method for generating and / or updating

Textures of background object models in a three-dimensional scene model of a surveillance scene with background objects, and a control device and a video surveillance system for performing the method and a computer program.

Video surveillance systems are used for camera-based control of relevant areas and usually include a plurality of surveillance cameras, which are set up in the relevant areas for recording surveillance scenes. The surveillance scenes can be designed, for example, as parking lots, intersections, streets, squares, but also as areas in buildings, factories, hospitals or the like. The image data streams recorded by the surveillance cameras are merged into a monitoring center where they are either automated or evaluated by monitoring personnel.

However, the work for the manual evaluation by the monitoring personnel is made more difficult by the fact that the image quality of the monitoring scenes shown can sometimes be classified as insufficient due to changes in lighting, the weather or contamination of the surveillance cameras. In order to simplify the work of the monitoring personnel and at the same time to improve the quality of surveillance, for example, the German patent application DE 10001252 A1 proposes a monitoring system which enables a more efficient working with a monitoring system through an object-oriented presentation. To do this, signals from the cameras used for each selected view are decomposed into objects and then passed to a display, adding artificial objects and deleting other objects.

Disclosure of the invention

In the context of the invention, a method for generating and / or updating textures of background object models in a three-dimensional scene model with the features of claim 1, a control device for carrying out the method with the features of claim 10, a video surveillance system according to claim 11 and a computer program the features of claim 12 proposed. Preferred or advantageous embodiments of the invention will become apparent from the dependent claims, the following description or from the figures.

The invention creates a possibility that surveillance scenes at least partially in a virtual reality or a partially virtual reality in

Form of a three-dimensional scene model can be displayed, with a particularly realistic representation of the surveillance scene can be achieved by generating and / or updating textures of background object models in the three-dimensional scene model. The fact that the surveillance scene is displayed on the one hand virtually, but on the other very realistically, it is for the

Surveillance personnel very simple and thus error-prone to switch between an actual observation of the surveillance scene and an observation of the virtual, three-dimensional scene model.

More generally, the method allows the mapping of a particular real

Surveillance scene with background objects on a three-dimensional scene model with background object models with realistic textures. As mentioned above, the surveillance scene can be streets, intersections, squares or areas in buildings, factories, prisons, hospitals, etc. The behind grand objects are preferably defined as static and / or quasi-static objects, which do not change or change only slowly and which are mapped to the background object models. Typical static objects are buildings, trees, panels, etc. The quasi-static objects include, for example, shadows, parked cars or the like. The static objects have a retention period in the

Monitoring scene of preferably more than several months, the quasi-static objects, however, preferably more than one or more days.

The three-dimensional scene model includes the background object models, each of which is formed as a three-dimensional model. For example, the three-dimensional scene model is "walkable", so that a user can move between the background object models in the three-dimensional scene model and / or change the view by adjusting the viewing direction or viewing angle / or an overlapping hierarchy (Z hierarchy) of

Background object models are stored.

The background object models and optionally the rest of the background have textures, with textures preferably being in the form of color, hatchings, patterns, and / or features of the surface of the background objects.

In one method step, a background image of the surveillance scene is formed from one or more camera images of the surveillance scene, wherein it is preferably provided that foreground objects or other interfering objects are masked out or suppressed. The background image can be formed in its thickness, ie in terms of the columns and rows of pixels, identical to the camera images. Alternatively, the background image is a section of one or more camera images. It is also possible that the background image has an arbitrary edge contour course, so that, for example, a background image represents exactly one background object.

In a further method step, the background image is projected onto the scene model. In this case, the background image is imaged in such a way that a pixel of a background object coincides with a corresponding model point of the background object model. The projection can also be done pixelwise in - A -

Take place in the form of a mapping rule, wherein preferably only the pixels are displayed, to which a corresponding model point is available.

After the background image is projected on the scene model or on the background object models, the textures of the

Created and / or updated background object models based on the projected background image. For this purpose, for example, image areas which are assigned in the correct position to the respective background object model after the projection are taken from the background image and utilized as a texture.

Optionally, the textures of the background object models are each stored with an orientation information, so that the textures in turn can be distributed on the background object models in the representation of the scene model on a monitor or the like, in the same manner as to the projection and projection.

In summary, the method allows to provide a three-dimensional scene model with realistic textures, wherein the textures can be updated at regular or irregular intervals.

In a preferred embodiment of the invention, the background image is through

Long-term observation, ie observation over several days, temporal filtering, thus for example by averaging, moving averaging, or by elimination of foreground objects formed. It is also possible to form the median of several camera images or to cut out known objects. In principle, all known methods for creating the background image can be used.

In a preferred implementation of the method, the projection of the background image onto the scene model takes place using the parameters of a camera model from the surveillance camera, from the perspective of which the background image is created. By using the parameters of the camera model, it is possible to perform the projection of a point in the coordinate system of the surveillance scene into the coordinate system of the camera image and vice versa. As an alternative to the camera model, it is also possible to use a look-up table which provides a corresponding point in the monitoring scene's coordinate system for each pixel in the camera image of the surveillance camera. By using an assignment rule between the surveillance scene and the camera image, it is possible to project the background image resulting from the camera image in the correct position and / or perspective corrected onto the scene model so that misallocations are minimized.

In an industry-oriented implementation of the method, the background image is optionally corrected in addition to distortions, which are unintentionally caused by aberrations in the surveillance camera system, e.g. as optical aberrations may have arisen, on the other hand, but also wanted distortions, which are introduced for example by the use of 360 ° - or fish eye cameras.

In a further preferred embodiment of the invention, the background image and / or image regions of the background image and / or pixels of the

Background image, in particular every pixel of the background image checked for occlusion with other static or quasi-static objects. In the event that the check reveals that the tested area is obscured by an interfering object, this pixel is discarded. Otherwise, the tested area will be used to create and / or update the textures.

In a further possible addition to the invention, a mutual occlusion of background object models is checked with the aid of a depth buffer, wherein pixels which are to be assigned to a background object model concealed in the region of the corresponding model point are discarded. For example, the depth buffer is based on a known Z-hierarchy from the rendering.

In a development of the invention, the textures are formed on the basis of a plurality of camera images, which originate from a common surveillance camera with a common viewing angle or from different surveillance cameras with different viewing angles onto the surveillance scene. Here, the camera images are projected from different angles in the manner described on the scene model in the correct position. After projection, pixels of different background images, which belong to a common texture point or a common texture of a background object model, are merged in each case. The Merging can be done, for example, by averaging. In a particularly preferred embodiment of the invention, for example, a color balance of the pixels to be merged takes place.

Optionally be supplemented, - in particular to gaps in one by the

Surveillance area formed monitoring scenes - texture information from other sources, such as aerial photographs taken.

In a particularly preferred embodiment of the method, the

Background object models with the textures in a display unit, such as a monitor or the like, of a video surveillance system, in particular as will be described below.

Another object of the invention relates to a video surveillance system which is connected to one or a plurality of surveillance cameras and / or connectable and which has a control device, characterized in that the control device circuitry and / or program technology for performing the method just described and / or it has been defined in the preceding claims is formed.

Particularly preferably, the video surveillance system is designed so that the described method preferably runs in the background at periodic intervals and thus keeps the textures up to date. A particular advantage of the video surveillance system is the fact that for the creation or

Updating the textures only the static and / or quasi-static scene background is considered. As a result, dynamic objects from the video images do not appear in the texture of the static geometry of the 3D model, which could otherwise result in a misrepresentation of the dynamic objects as a texture on the background object models, eg flat on the road or on walls. The dynamic objects, on the other hand, can be displayed separately - either as a real image or as a virtual representation - in the scene model, which leads to a credible or realistic visualization. A final object of the invention relates to a computer program with program code means for carrying out all the steps of the described method when the program is executed on a computer and / or a video surveillance system.

Brief description of the drawings

Further features, advantages and effects of the invention will become apparent from the following description of a preferred embodiment of the invention and the accompanying drawings. Showing:

FIG. 1 is a flow chart illustrating a first embodiment of the method according to the invention;

Figure 2 is a block diagram of a video surveillance system for performing the

Method according to FIG. 1

Embodiment (s) of the invention

FIG. 1 shows, in a schematic flowchart, the sequence of a method for generating and / or updating textures of background object models in a three-dimensional scene model as an exemplary embodiment of the invention.

As current input information, one or more video images 1, from

Surveillance cameras are 10 (Figure 2), used. These video images 1 are converted in a first method step 2 into a background image with background pixels. The transfer is accomplished by methods known in image processing, such as averaging or median formation of multiple video frames 1, clipping known objects, long-term observation, or the like.

By this method step, a background image is generated, which has as active pixels only background pixels from the one or more video images 1 and optionally disabled pixels, which are set to the positions of the video image 1, where a disturber object or foreground object is shown. In a second method step 3, the background image generated in this way is projected onto a scene model. The scene model is designed as a three-dimensional scene model and has a multiplicity of background object models, for example representative of buildings, furniture, streets or other stationary objects. Within the scope of method step 3, with the aid of the parameters of the camera model, the

Surveillance cameras that have supplied the background image underlying video image 1, which projects pixels of the background image in the image coordinate system to the respective corresponding point of the three-dimensional scene model. Optionally, in addition to the distortion distortions, such as distortions or the like are corrected.

In a third method step 4, a pixel-by-pixel check for occlusion from the camera's view takes place with the aid of a depth buffer. Here, it is tested whether a pixel of the background image, which was projected by the method step 3 on a background object model, by another background object model and / or a real, e.g. dynamic object in the current camera view is obscured. In the event that the test assesses the inspected pixel as hidden, it is discarded and no longer used. Otherwise, the pixel, ie the projected video pixel or background pixel, is used for creating and / or updating the textures.

In a fourth method step 5, the textures 6 are created and output on the basis of the transferred background pixels. As additional measures it can be provided that several pixels of different background images, which are arranged overlapping at least in sections after the projection and thus relate to the same regions of the same background object models, are merged into a common background pixel. Here, e.g. also a color match done. In particular, any remaining gaps in the scene model may be filled by static textures, e.g. taken from aerial photographs.

FIG. 2 shows a video surveillance system 100 which is designed to carry out the method described in FIG. The video surveillance system is connected to a plurality of security cameras 10 wirelessly or wired signal technology and is designed for example as a computer system. The Surveillance cameras 10 are directed to relevant areas which show surveillance scenes in the form of squares, intersections or the like.

The image data streams of the surveillance cameras 10 are transferred to a background module 20, which is designed to carry out the first method step 2 in FIG. The background image (s) generated are forwarded to a projection module 30, which is designed to carry out the second method step 3. To control for occlusion, the projected background images are passed to a masking module 40, which is implemented for carrying out the third method step 4. In a texture module 50, the textures 6 are created or updated on the basis of the checked background images and forwarded to a texture memory 60.

Based on the stored data and the three-dimensional scene model is displayed on a display unit 70, such. a virtual representation of a monitor

Surveillance scene presented with background object models with real textures. In this virtual representation, real objects, such as e.g. Dynamic objects in the surveillance scene are displayed in the correct position and close to reality.

Claims

claims 1. A method for generating and / or updating textures (6) of background object models in a three-dimensional scene model of a Surveillance scene with background objects, wherein a background image of the surveillance scene is formed from one or more camera images (1) of the surveillance scene (2), where the background image is projected onto the scene model (3) and wherein textures of the background object models are generated and / or updated based on the projected background image (5). 2. The method according to claim 1, characterized in that the background image is formed by long-term observation, filtering and / or by elimination of foreground objects. 3. The method according to claim 1 or 2, characterized in that the projection of the background image is carried out using a camera model. 4. The method according to any one of the preceding claims, characterized in that the background image is projected in the correct position and / or perspective corrected projected onto the scene model. 5. The method according to claim 3 or 4, characterized in that the background image is corrected for distortion. 6. The method according to any one of the preceding claims, characterized in that to the background image and / or an image area of the background image and / or a pixel of the background image corresponding area of a background object model is checked for obstruction by other background object models (4). 7. The method according to any one of the preceding claims, characterized in that the textures (6) are formed on the basis of a plurality of camera images (1), which originate from different angles on the surveillance scene. A method according to claim 7, characterized in that pixels of different background images belonging to a common texture point or a common texture of a background object model are merged. 9. The method according to any one of the preceding claims, characterized in that the background object models with the textures in a display unit of a Video surveillance system (100) are shown. 10. Control device (100), characterized in that the control device (100) circuitry and / or program technology for performing the method according to one of the preceding claims is formed. 11. A video surveillance system connected to or connectable to one or a plurality of surveillance cameras (10), characterized in that the video surveillance system comprises a control device (100) according to claim 10. A computer program with program code means for carrying out all the steps of the method according to any one of claims 1 to 9, when the program is executed on a computer and / or a control device according to claim 10 and / or a video surveillance system according to claim 11.