Camera digital gimbal system
View Patent ↗A camera system has at least a first and a second camera which are attached to a common platform, each camera having a lens and an image sensor. Active areas of the image sensors of the first and second camera are determined. Image data only gathered from active areas of image sensors establish a panoramic image, which may be called an extended image space. Position sensors allow a processor to determine a location and a pose of the camera system. Coordinates of the extended image space are associated with a position of an object in physical space relative to a pointing direction of the camera system. A screen displays image data of a window of preset size around the image of the object in extended image space. A neural network is used to train the camera system to predict a position of an object based on camera rotations.
1. A processor implemented method, comprising:
determining an extended image space of active areas of image sensors in at least a first camera and a second camera in a fixed position relative to the first camera in a common housing, active areas of image sensors of two or more cameras being determined by non-overlap of image areas generated by the image sensors of the at least first and second cameras, the extended image space representing a substantially registered image;
recording a sequence of video frames with the active areas of the image sensors of the at least first and second cameras;
collecting motion sensor data from one or more motion sensors associated with the common housing, wherein a movement of the common housing comprises at least one of a pitch, and a yaw;
determining coordinates of a point in extended image space with the common housing in a first pointing direction of the common housing and determining modified coordinates of the point in extended image space with the common housing in a second pointing direction of the common housing based on data generated by the one or more motion sensors;
capturing image data determined by a window being smaller than a size of the extended image space around the modified coordinates of the at least one point in the extended image space in the second position; and
displaying on the screen an image determined by the captured image data determined by the window.
2. The method of claim 1 , wherein the movement of the common housing comprises at least one of a pitch, a roll and a yaw.
3. The method of claim 2 , further comprising:
compensating the roll by scanline jumping based on the motion sensor data collected from the one or more motion sensors.
4. The method of claim 1 , further comprising:
determining the coordinates of the point associated with the first pointing direction based on a center of an object in the extended image space.
5. The method of claim 4 , further comprising:
tracking the object in the extended image space in extended image space by an object tracker and determining the modified coordinates based on coordinates of the tracked object in extended image space with the common housing in the second position, wherein a center of the tracked object is a center of an e-gimbal window.
6. The method of claim 4 , further comprising:
training a convolutional neural network to perform object tracking in the extended image space.
7. The method of claim 1 , wherein the first and second cameras are identical and have a curved image sensor.
8. The method of claim 1 , wherein the first and second cameras are self-aligning.
9. The method of claim 1 , further comprising:
displaying the captured image data at a frame rate of at least 10 frames per second.
10. The method of claim 1 , further comprising:
associating with a neural network data generated by the one or more motion sensors with modified coordinates in the extended image space.
11. The method of claim 1 , further comprising:
predicting by a neural network one or more rotation corrective instructions based on one or more inputs from the positional sensors.
12. A camera system, comprising:
one or more storage devices enabled to store data, including instructions;
a screen to display an image;
one or more processors, enabled to retrieve instructions from the one or more storage devices, the instructions when executed by the one or more processors cause to carry out computer functions, comprising:
determining an extended image space of active areas of image sensors in at least a first camera and a second camera in a fixed position relative to the first camera in a common housing, active areas of image sensors of two or more cameras being determined by non-overlap of image areas generated by the image sensors of the at least first and second cameras, the extended image space representing a substantially registered image;
recording a sequence of video frames with the active areas of the image sensors of the at least first and second cameras;
collecting motion sensor data from one or more motion sensors associated with the common housing, wherein a movement of the common housing comprises at least one of a pitch, and a yaw;
determining coordinates of a point in extended image space with the common housing in a first pointing direction of the common housing and determining modified coordinates of the point in extended image space with the common housing in a second pointing direction of the common housing based on data generated by the one or more motion sensors;
capturing image data determined by a window being smaller than a size of the extended image space around the modified coordinates of the at least one point in the extended image space in the second position; and
displaying on the screen an image determined by the captured image data determined by the window.
13. The camera system of claim 12 , wherein the movement of the common housing comprises at least one of a pitch, a roll and a yaw.
14. The camera system of claim 13 , further comprising to carry out the computer function:
compensating the roll by scanline jumping based on the motion sensor data collected from the one or more motion sensors.
15. The camera system of claim 12 , further comprising:
determining the coordinates of the point associated with the first pointing direction based on a center of an object in the extended image space.
16. The camera system of claim 15 , further comprising:
tracking the object in the extended image space by an object tracker and determining the modified coordinates based on coordinates of the tracked abject in extended image space with the common housing in the second position.
17. The camera system of claim 15 , further comprising:
training a convolutional neural network to perform object tracking in the extended image space.
18. The camera system of claim 12 , wherein the first and second cameras are identical and have a curved image sensor.
19. The camera system of claim 12 , further comprising:
associating with a neural network data generated by the one or more motion sensors with modified coordinates in the extended image space.
20. The camera system of claim 12 , further comprising:
predicting by a neural network one or more rotation corrective instructions based on one or more inputs from the positional sensors.