The present invention relates to a camera and a corresponding method for selecting an object to be recorded.
Frequently, a camera is used to record moved scenes freehand. If an object that moves independently is located in the scene to be recorded, the camera must be tracked corresponding to the movement of the object. Even professional camera operators have only limited success with this tracking or the associated deliberate panning movement of the camera. For the stable tracking of the camera, the movement of the object should be measured or ascertained. For this, it is necessary for the moved object to be segmented in the successively recorded images. Apart from the desired object to be recorded, other moved objects which have similar or different movements or a static background may also be segmented from the analysis of the recorded images. Since a plurality of objects may be discovered in the recorded images, additional information is therefore required for the selection of the desired object. This information may be input, for example, using an appropriate user menu; however, this results in the loss of the desired tracking in the case of fast moving objects.
An image stabilization system and a corresponding method for a video camera are discussed in unexamined patent application US 2005/028070 A1. The described image stabilization system for automatic tracking includes a camera having an image capturing device which records a video image. The camera has a selectively adjustable image area. A processing unit is operatively coupled to the camera and receives signals which include the image area of the camera and the recorded images. In the captured images, the processing unit automatically tracks the movements of at least one object and determines a stabilization setting for the recorded video image as a function of an inadvertent change in the image area of the camera during a time interval between the recording of a first image and the recording of a second image. The determination of the stabilization setting is based on the signals which indicate the image area and an analysis of the first and second image.
The camera according to the exemplary embodiments and/or exemplary methods of the present invention having the features described herein has the advantage over the related art that an evaluation and control unit ascertains a panning movement of the camera and compares it with ascertained movements of objects which are represented in the images captured by an image capturing device, the evaluation and control unit determining at least one object as an object to be recorded, the ascertained movement of which is most consistent with the ascertained panning movement of the camera, and the evaluation and control unit storing an image section of the image captured by the image capturing device in a storage unit, the image section representing the at least one object to be recorded.
The method according to the present invention for selecting an object to be recorded having the features described herein has the advantage over the related art that a panning movement of the camera is ascertained during the image recording, movements of objects being ascertained in the captured images and compared with the camera movement, at least one object being determined as an object to be recorded, the ascertained movement of which is most consistent with the ascertained panning movement of the camera, and an image section of the captured images being stored, the image section representing the at least one object to be recorded.
The camera according to the present invention and the method according to the present invention make it possible to automatically hold the desired object to be recorded, which the user advantageously tracks, for example, by a deliberate panning movement of the camera, in the stored image section. Since by manually tracking the camera, the user attempts to hold the desired object nearly constantly in the image section to be recorded, the desired object has nearly the same movement parameters as the camera, so that the camera movement may be used advantageously as a selection criterion. Since the manual tracking is insufficiently precise and the image is often blurred during tracking, it is of particular benefit to stabilize the image section having the desired object to be recorded by using appropriate known stabilization methods.
The measures and refinements listed in the further description herein make it possible to achieve advantageous improvements on the camera specified herein and on the method specified herein for selecting an object to be recorded.
It is advantageous in particular that the evaluation and control unit digitally evaluates the recorded image sequence for ascertaining an optical flow, the evaluation and control unit evaluating the ascertained optical flow in order to generate information concerning the panning movement of the camera and/or concerning the movements of the objects in the images. The evaluation of the optical flow makes it possible to recognize and segment image areas having the same direction of movement in the captured image. To generate the information concerning the camera movement, the evaluation and control unit may, for example, evaluate the ascertained optical flow of a recognized image background. Alternatively, the deliberate panning movement of the camera may be detected by a sensor unit which transfers information concerning the camera movement to the evaluation and control unit for evaluation. In digital image processing technology, “optical flow” is understood to be a vector field which specifies for each pixel the two-dimensional movement of the corresponding pixel in an image sequence.
In one embodiment of the present invention, the resolution of the images captured by the image capturing device is significantly greater than the resolution of the images stored in a storage unit. This advantageously makes it easier to recognize moved objects which carry out the same movement as the camera.
In another embodiment of the present invention, the evaluation and control unit establishes a focus of the at least one object to be recorded as a central point of the image section to be recorded and scales the size of the image section to be recorded to the size used in the storage unit. The central point and the size of the image section to be stored are calculated using the distribution of the selected objects.
In another embodiment of the present invention, a common focus is established as a central point of the image section to be recorded and/or an object is filtered out using a temporal filter if a plurality of objects is determined having the same direction of movement as objects to be recorded. This makes it possible to use the ascertained central point of the selected objects also for stabilizing the image section.
Exemplary embodiments of the present invention are represented in the drawings and are elucidated in greater detail in the following description.
The function of the camera according to the present invention is described in the following with reference to
As is apparent from
For generating information concerning the panning movement of the camera and/or the movements of the objects in the captured images, evaluation and control unit 10 evaluates an optical flow which is ascertained by evaluation and control unit 10 using a digital evaluation of the recorded image sequence. Evaluation and control unit 10 is thus able to evaluate, for example, the optical flow of a recognized image background in order to generate information concerning the panning movement of camera 1. In addition or alternatively, evaluation and control unit 10 is able to receive information concerning the panning movement of camera 1 from sensor unit 40 which detects the panning movement of camera 1 using corresponding sensors. Sensor unit 40 may, for example, have inertial sensors which provide information concerning the panning movement of camera 1 via the detection of yaw rates. To make tracking of a desired object easier, the resolution of the images captured by image capturing device 20 is selected to be significantly greater than the resolution of the image sections to be stored in storage unit 30. In order to establish the image sections to be recorded or to be stored, evaluation and control unit 10 determines a focus of the at least one object to be recorded as a central point of the image section to be recorded, evaluation and control unit 10 scaling the size of the image section to be recorded to the size used in storage unit 30.
Camera 1 according to the present invention thus makes possible a stabilized tracking of a desired object by comparing the panning movement of camera 1 with the movement of objects in the image, it being possible for the camera movement to be measured using either the inertial movements or the optical flow. The objects may be segmented using the ascertained optical flow. The movements of the objects are measured for the purpose of stable tracking. From the analysis of the dense optical flow, evaluation and control unit 10 ascertains all objects in the captured images having different movements and the static background. However, since a plurality of objects having similar or different movements may be discovered in the captured images, the selection of the desired object is made by comparison with the panning movement of camera 1. If there is a plurality of objects having a comparable movement, the common focus of the selected objects may be used as a central point of the image section to be recorded and for stabilization and/or an object may be filtered out using a temporal filter. Alternatively, the desired object to be recorded could be selected in this case using operating unit 50.
In summary, camera 1 according to the present invention contains an image capturing device 20 having a significantly greater resolution than is necessary for recording image sequences. In these image sequences, the objects are segmented using their similar image flow, the desired movement of the object to be tracked being established via the panning movement of camera 1. To that end, the panning movement is determined using either inertial sensors (yaw rate) or using the image flow of the background (analysis of the intrinsic movement). If one or a plurality of objects moves according to the panning movement of camera 1, the common focus is selected as the central point of the image section to be recorded, i.e., the active image section is positioned in such a way that the object retains its absolute image coordinates, or the desired object may be filtered out using a temporal filter. If necessary, the size is tracked using automatic image scaling, making it possible to capture all objects of the desired direction of movement.
Number | Date | Country | Kind |
---|---|---|---|
10 2008 042 562.1 | Oct 2008 | DE | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/EP2009/062193 | 9/21/2009 | WO | 00 | 8/8/2011 |