Motion-based video stabilization

A motion sensor-based video stabilization method, comprising:

• capturing a plurality of sequential images by a video capture device, the video capture device having image capture parameters wherein each of the plurality of sequential images is associated with values corresponding to the image capture parameters at the time each of the images was captured;
• capturing, at the video capture device, motion information for each of the plurality of sequential images, wherein the motion data for each image in the plurality of sequential images is captured at approximately the same time as each image was captured;
• estimating an unwanted motion for each of the plurality of sequential images based on each image's motion information;
• estimating intrinsic data from the image capture parameter values, wherein each item of intrinsic data represents a mapping from a three-dimensional space to a two-dimensional space associated with a plane of the video capture device in the absence of motion, and wherein a respective item of intrinsic data is estimated for each of the plurality of sequential images;
• applying a transform to each of the plurality of sequential images to substantially remove the estimated unwanted motion, wherein the transform applied to each of the plurality of sequential images is based on each image's image capture parameter values and motion information, and wherein the transform includes applying the intrinsic data to respective images;
• and
• storing each of the transformed plurality of sequential images in a memory.

A non-transitory storage device having instructions stored thereon for causing a programmable control device to perform the following:

• capturing, by a video capture device, a video sequence having a plurality of sequential images, each of the plurality of sequential images associated with one or more image capture parameter values representing operational parameters of the video capture device;
• capturing, at the video capture device, motion data for each of the plurality of sequential images;
• estimating unwanted motion of the video capture device for each of the plurality of sequential images based on the motion data;
• estimating intrinsic data from the image capture parameter values, wherein each item of intrinsic data represents a mapping from a three-dimensional space to a two-dimensional space associated with a plane of the video capture device in the absence of motion, and wherein a respective item of intrinsic data is estimated for each of the plurality of sequential images;
• modifying each of the plurality of sequential images to remove the estimated unwanted motion based on the motion data and the image capture parameter values associated with each of the plurality of sequential images, wherein the modifying includes applying the intrinsic data to respective images;
• and storing each of the modified plurality of sequential images in a memory.

An electronic device, comprising:

• a video capture sensor;
• a motion sensor;
• a programmable control device communicatively coupled to the memory;
• and
• a memory communicatively coupled to the video capture circuit, the motion sensor, and the programmable control device, the memory having computer program code stored therein for causing the programmable control device to—
• capture a plurality of sequential images from the video capture sensor, the electronic device having image capture parameters wherein each of the plurality of sequential images is associated with values corresponding to the image capture parameters at the time each of the images was captured;
• capture motion information at the motion sensor for each of the plurality of sequential images, wherein the motion data for each image in the plurality of sequential images is captured at approximately the same time as each image was captured;
• estimate an unwanted motion for each of the plurality of sequential images based on each image's motion information;
• estimating intrinsic data from the image capture parameter values, wherein each item of intrinsic data represents a mapping from a three-dimensional space to a two-dimensional space associated with a plane of the video capture device in the absence of motion, and wherein a respective item of intrinsic data is estimated for each of the plurality of sequential images;
• apply a transform to each of the plurality of sequential images to substantially remove the estimated unwanted motion, wherein the transform applied to each of the plurality of sequential images is based on each image's image capture parameter values and motion information, and wherein the transform includes applying the intrinsic data to respective images;
• and
• store each of the transformed plurality of sequential images in the memory.