IPC분류정보
국가/구분 |
United States(US) Patent
등록
|
국제특허분류(IPC7판) |
|
출원번호 |
US-0312862
(2011-12-06)
|
등록번호 |
US-8890962
(2014-11-18)
|
발명자
/ 주소 |
- Proca, Adrian
- Jacobs, William
|
출원인 / 주소 |
|
대리인 / 주소 |
Patterson + Sheridan, LLP
|
인용정보 |
피인용 횟수 :
2 인용 특허 :
2 |
초록
▼
Systems and methods for performing digital image stabilization implemented in a digital camera. The digital camera includes a gyroscope to measure motion of the digital camera and processes the signals from the gyroscope to track the total displacement of an image sensor over a series of frames of v
Systems and methods for performing digital image stabilization implemented in a digital camera. The digital camera includes a gyroscope to measure motion of the digital camera and processes the signals from the gyroscope to track the total displacement of an image sensor over a series of frames of video. The algorithm implemented by the digital camera includes a processing block for correcting a DC bias in the signals from the gyroscope, a filter for attenuating the signals during periods of high acceleration, a processing block for detecting the start of a panning motion, and a processing block for quickly retracing the digital image stabilization correction back to the center of the image sensor during a panning motion.
대표청구항
▼
1. A method, comprising: receiving motion feedback data via a motion sensor configured to detect motion associated with a digital camera;determining a DC (direct current) level associated with the motion sensor that corresponds to the digital camera having zero motion;modifying the motion feedback d
1. A method, comprising: receiving motion feedback data via a motion sensor configured to detect motion associated with a digital camera;determining a DC (direct current) level associated with the motion sensor that corresponds to the digital camera having zero motion;modifying the motion feedback data based on the DC level;attenuating the motion feedback data based on a measure of acceleration of the digital camera;identifying a plurality of active pixel locations based on the modified motion feedback data that comprise a subset of all pixel locations associated with an image sensor of the digital camera, wherein each active pixel location is offset from a corresponding pixel location associated with the image sensor, and each corresponding pixel location is associated with a zero offset; andgenerating a digital image based on the plurality of active pixel locations. 2. The method of claim 1, wherein the motion sensor comprises a gyroscope, and the motion feedback data comprises rotational velocity data corresponding to one or more axes of the digital camera. 3. The method of claim 2, wherein the rotational velocity data corresponds to an x-axis, a y-axis and a z-axis of the digital camera. 4. The method of claim 2, wherein the motion sensor further comprises an accelerometer, and the motion feedback data further comprises translational velocity data corresponding to dimensions associated with the one or more axes. 5. The method of claim 1, wherein determining the DC level comprises: comparing the DC level to the motion feedback data to determine whether the motion feedback data is greater than the DC level; andif the motion feedback data is greater than the DC level, then incrementing the DC level by a DC level adaptation step, orif the motion feedback data is less than or equal to the DC level, then decrementing the DC level by the DC level adaptation step. 6. The method of claim 5, further comprising calculating an initial DC level based on a temperature associated with the motion sensor. 7. The method of claim 6, wherein the initial DC level is calculated using a linear equation having a slope coefficient and an offset coefficient that are determined during a calibration routine performed during the manufacture of the digital camera. 8. The method of claim 1, wherein attenuating the motion feedback data comprises: differentiating the modified motion feedback data to determine the acceleration of the digital camera;determining an attenuation coefficient based on the acceleration of the digital camera; andmultiplying the modified motion feedback data by the attenuation coefficient. 9. The method of claim 8, wherein determining the attenuation coefficient is performed by referencing a look-up table (LUT), and wherein an index into the LUT is generated based on the acceleration of the digital camera. 10. The method of claim 1, further comprising detecting that the digital camera is undergoing a panning motion. 11. The method of claim 10, wherein detecting that the digital camera is undergoing a panning motion comprises: determining that the displacement of the digital camera based on the modified motion feedback data is greater than a first threshold value; andtriggering a pan detection flag to indicate that the digital camera is in the panning motion. 12. The method of claim 11, wherein detecting that the digital camera is undergoing a panning motion further comprises, in addition to determining that the displacement of the digital camera based on the modified motion feedback data is greater than the first threshold value, determining that the velocity of the digital camera indicated by the modified motion feedback data is greater than a second threshold value. 13. The method of claim 12, wherein detecting that the digital camera is undergoing a panning motion further comprises, in addition to determining that the velocity of the digital camera indicated by the modified motion feedback data is greater than the second threshold value, determining that a number of consecutive samples of the modified motion feedback data having a uniform direction is greater than a third threshold value. 14. The method of claim 11, wherein triggering a pan detection flag comprises setting a bit in a register of the digital camera. 15. The method of claim 10, further comprising, in response to detecting that the digital camera is undergoing the panning motion, activating a fast zero recovery with alpha blending operation that causes the offset associated with each active pixel location to approach a zero offset. 16. The method of claim 15, wherein the fast zero recovery with alpha blending operation comprises: calculating a displacement vector at the start of the panning motion that represents, for each active pixel location, the offset of the active pixel location from the corresponding pixel location; andfor each sampling of the motion feedback data: generating a retrace vector by subtracting a fixed value from each component of the displacement vector, andblending the displacement vector with the retrace vector to generate a stabilization vector that, when applied to the one or more active pixel locations, causes the offset associated with each active pixel location to approach a zero offset. 17. The method of claim 16, wherein blending the displacement vector with the retrace vector comprises: calculating a decay rate based on an elapsed time from the start of the panning motion; andsumming the product of the displacement vector and the quantity one minus the decay rate with the product of the retrace vector and the decay rate. 18. The method of claim 1, wherein the image sensor is a complementary metal oxide semiconductor (CMOS) image sensor. 19. An digital camera, comprising: a motion sensor configured to detect motion associated with the digital camera;an image sensor; andlogic that, when executed, performs an operation, the operation comprising: receiving motion feedback data via the motion sensor;determining a DC (direct current) level associated with the motion sensor that corresponds to the digital camera having zero motion;modifying the motion feedback data based on the DC level;attenuating the motion feedback data based on a measure of acceleration of the digital camera;identifying a plurality of active pixel locations based on the modified motion feedback data that comprise a subset of all pixel locations associated with the image sensor, wherein each active pixel location is offset from a corresponding pixel location associated with the image sensor, and each corresponding pixel location is associated with a zero offset; andgenerating a digital image based on the plurality of active pixel locations. 20. A non-transitory computer-readable medium containing computer program code that, when executed, performs an operation comprising: receiving motion feedback data via a motion sensor configured to detect motion associated with a digital camera;determining a DC (direct current) level associated with the motion sensor that corresponds to the digital camera having zero motion;modifying the motion feedback data based on the DC level;attenuating the motion feedback data based on a measure of acceleration of the digital camera;identifying a plurality of active pixel locations based on the modified motion feedback data that comprise a subset of all pixel locations associated with an image sensor of the digital camera, wherein each active pixel location is offset from a corresponding pixel location associated with the image sensor, and each corresponding pixel location is associated with a zero offset; andgenerating a digital image based on the plurality of active pixel locations.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.