Apparatus and methods for rolling shutter compensation for multi-camera systems

• US 10,560,633 B2
• Filed: 07/20/2018
• Issued: 02/11/2020
• Est. Priority Date: 10/29/2015
• Status: Active Grant

• Current Assignee: GoPro, Inc.
• Original Assignee: GoPro, Inc.
• Inventors: Meler, Antoine
• Primary Examiner(s): Czekaj, Dave
• Assistant Examiner(s): Brumfield, Shanika M
• Application Number:
  US16/041,380
• Publication Number:
  US 20180332206A1
• Time in Patent Office: 571 Days
• Field of Search: None
• CPC Class Codes: G02B 13/06 Panoramic objectives; So-called "sky lenses" including panoramic objectives having reflecting surfaces
• CPC Class Codes: G06T 3/047 Fisheye or wide-angle transformations
• CPC Class Codes: G06T 3/4038 Image mosaicing, e.g. composing plane images from plane sub-images
• CPC Class Codes: H04N 23/689 Motion occurring during a rolling shutter mode
• CPC Class Codes: H04N 23/698 for achieving an enlarged field of view, e.g. panoramic image capture
• CPC Class Codes: H04N 23/90 Arrangement of cameras or camera modules, e.g. multiple cameras in TV studios or sports stadiums
• CPC Class Codes: H04N 25/41 Extracting pixel data from a plurality of image sensors simultaneously picking up an image, e.g. for increasing the field of view by combining the outputs of a plurality of sensors
• CPC Class Codes: H04N 25/531 by controlling rolling shutters in CMOS SSIS
• CPC Class Codes: H04N 25/76 Addressed sensors, e.g. MOS or CMOS sensors
• CPC Class Codes: H04N 7/181 for receiving images from a plurality of remote sources

Abstract

Images may be obtained using a moving camera comprised of two or more rigidly mounted image sensors. Camera motion may change camera orientation when different portions of an image are captured. Pixel acquisition time may be determined based on image exposure duration and position of the pixel in the image array (pixel row index). Orientation of the sensor may at the pixel acquisition time instance may be determined. Image transformation may be performed wherein a given portion of the image may be associated with a respective transformation characterized by the corrected sensor orientation. In some implementations of panoramic image acquisition, multiple source images may be transformed to, e.g., equirectangular plane, using sensor orientation that is corrected for the time of pixel acquisition. Use of orientation correction may improve quality of stitching by, e.g., reducing contrast of border areas between portions of the transformed image obtained by different image sensors.

Claims

1. A computerized system configured to obtain composite images, the system comprising:

a processor adapted to execute a plurality of computer instructions; and

non-transitory storage medium including the plurality of the computer instructions which, when executed by the processor, cause the processor to;

obtain component images, the component images including a first component image captured by a first imaging sensor and a second component image captured by a second imaging sensor on a row-by-row basis, the row-by-row capture of the component images resulting in capture of individual rows of the component images at different acquisition times;

generate a first composite image by performing a first transformation operation on the component images, wherein the first component image includes an overlapping area that corresponds to an overlap between the first component image and the second component image;

for at least some pixels in the overlapping area of the first composite image, determine corresponding rows in at least one of the component images such that for a first set of pixels in the first composite image a first row of the first component image is determined to be corresponding;

determine acquisition times of the at least one of the component images associated with row locations corresponding to the pixels in the first composite image such that a first acquisition time is determined for the first row of the first component image;

determine orientation of the first imaging sensor based on the acquisition times and orientation information of the first imaging sensor such that a first orientation of the first imaging sensor is determined for capture of the first row of the first component image by the first imaging sensor based on the first acquisition time; and

perform a second transformation operation on the first component image based on the first orientation of the first imaging sensor to generate a second composite image such that the second transformation operation compensates for the row-by-row basis capture of the first component image and the orientation of the first imaging sensor.

2. The system of claim 1, wherein:

the first composite image is characterized by a first performance metric and the second composite image is characterized by a second performance metric; and

the second transformation operation is configured to improve the second performance metric relative the first performance metric.

3. The system of claim 1, wherein:

the first imaging sensor is configured to obtain the first component image at a first time instance; and

the orientation information is obtained based on analysis of the first component image and a third component image obtained by the first imaging sensor at a second time instance prior to the first time instance.

4. The system of claim 1, further comprising:

a sensor interface configured to obtain the orientation information.

5. The system of claim 4, wherein the sensor interface is configured to be coupled to an orientation sensor, the orientation sensor comprising one or more of a gyroscope and/or accelerometer.

6. The system of claim 1, further comprising:

a support structure configured to support the first imaging sensor, and to support the second imaging sensor at a given orientation relative to the first imaging sensor; and

an orientation sensor configured to obtain an orientation of the support structure;

wherein the orientation information is configured based on an orientation of the first imaging sensor and an orientation of the second imaging sensor relative to the support structure, and the orientation of the support structure.

7. The system of claim 1, wherein:

determination of the corresponding rows in the component images is configured based on an iteration process such that for a given iteration, a first instance of the corresponding first row of the first component image is determined based on a second instance of the corresponding first row of the first component image determined at another iteration preceding the given iteration.

8. The system of claim 1, further comprising:

a first lens characterized by a first lens projection, the first lens configured to provide light waves to the first imaging sensor; and

wherein the first transformation operation is configured based on an inverse of the first lens projection.

9. The system of claim 8, further comprising:

a second lens characterized by a second lens projection, the second lens configured to provide light waves to the second imaging sensor; and

wherein the first transformation operation is further configured based on an inverse of the second lens projection.

10. The system of claim 9, wherein

the first component image is configured based on the first lens projection and the first composite image is configured in an equirectangular projection;

and

the first transformation is configured to transform the first component image and the second component image into equirectangular plane.

11. The system of claim 10, wherein:

the first lens projection is characterized by a rectilinear transformation.

12. The system of claim 10, wherein:

the first lens projection is characterized by a wide angle fisheye transformation such that a first lens field of view is characterized by a first cardinal dimension configured in excess of 120 degrees of arc and a second cardinal dimension configured in excess of 90 degrees of arc.

13. The system of claim 12, wherein:

the first imaging sensor is disposed in a plane configured perpendicular to plane of the second imaging sensor; and

the second lens projection is configured based on the wide angle fisheye transformation such that a second lens field of view is characterized by the second cardinal dimension and the first cardinal dimension.

14. The system of claim 13, wherein the first lens field of view and the second lens field of view overlap such that pixels of the first component image corresponding to the overlapping portion and pixels of the second component image corresponding to the overlapping portion are transformed into pixels of the overlapping area within the first composite image.

15. The system of claim 1, wherein the first imaging sensor comprises an array of rows of photosites configured based on a CMOS imaging process wherein time of acquiring content of a given row of photosites of the array is delayed relative to a time instance of acquiring content of another row of photosites of the array.

16. A computer implemented method of improving quality of a digital panorama image obtained based on stitching of source digital images, the method comprising:

for a portion of the digital panorama image, determining a corresponding location in at least one of the source digital images, the at least one of the source digital images acquired by an image sensor on a row-by-row basis such that individual row of the at least one of the source digital images are acquired at different acquisition times;

determining an acquisition time at which the image sensor acquired the corresponding location in the at least one of the source digital images;

determining an orientation of the image sensor during the acquisition time; and

performing a stitching operation configured to transform a portion of the at least one source digital image from a coordinate representation of the image sensor into a coordinate representation of the digital panorama image, the transform configured based on the orientation of the image sensor during the acquisition time to compensate for acquisition of the individual rows of the at least one of the source digital images at different acquisition times and the orientation of the image sensor.

17. The method of claim 16, wherein:

the image sensor is configured to change its orientation during the acquisition of the at least one source digital image; and

a nominal sensor orientation is maintained at a given orientation during the acquisition of the at least one source digital image.

18. The method of claim 16, wherein:

the at least one source digital image includes multiple rows of pixels;

the image sensor comprises a CMOS image sensor configured to acquire a given row of pixels of the multiple rows of pixels at a given time;

the digital panorama image includes multiple pixels;

the portion of the digital panorama image corresponds to at least one pixel of the multiple pixels;

the portion of the at least one source digital image corresponds to a row of the multiple rows of pixels; and

the corresponding location corresponds to an index of the row within the multiple rows of pixels of the at least one source digital image.

19. The method of claim 16, wherein the portion of the at least one source digital image is configured to be captured by the image sensor at a first time instance, and another portion of the source digital image is configured to be captured by the image sensor at a second time instance, the second time instance spaced by a delay interval from the first time instance, the delay interval is selected from a range between 5 microseconds and 1000 microseconds.

20. A non-transitory computer readable medium comprising a plurality of computer instructions configured to, when executed by a processor, produce a composite panoramic image in a panoramic image plane from first and second source images comprised of rows of pixels, the first and second source images captured on a row-by-row basis such that individual rows of the first and second source images are captured at different acquisition times, by at least:

for a location in the panoramic image plane, determine corresponding rows in at least one of the first source image or the second source image such that a corresponding first row of pixels in the first source image is determined;

determine acquisition times of at least one of the first source image or the second source image associated with the corresponding rows such that a first acquisition time is determined for the first row of pixels in the first source image;

determine sensor orientations based on the acquisition times and orientation information such that a first sensor orientation is determined for a first sensor that captured the first source image at the first acquisition time; and

perform a transformation operation on the first source image based on the first sensor orientation to obtain the composite panoramic image such that the transformation operation compensates for the row-by-row basis capture of the first component image by the first sensor and the first sensor orientation of the first sensor.