Motion vector detection apparatus and motion vector detection method to accurately detect a motion vector from a captured image utilizing a degree of reliability information
Provided is a first motion vector detection unit that detects a motion vector on a basis of self-motion data and depth image data. A second motion vector detection unit that detects a motion vector on a basis of captured image data. A motion vector merging unit that merges the motion vector detected by the first motion vector detection unit with the motion vector detected by the second motion vector detection unit.
1. A motion vector detection apparatus, comprising:
a motion sensor configured to acquire self-motion data, depth image data, and captured image data; and
circuitry configured to:
detect a first motion vector based on the self-motion data and the depth image data;
detect a second motion vector based on the captured image data;
calculate a first predicted error of the first motion vector and a second predicted error of the second motion vector;
calculate a first degree of reliability and a second degree of reliability based on the first predicted error and the second predicted error, respectively; and
merge the first motion vector with the second motion vector based on the first degree of reliability and the second degree of reliability.
2. The motion vector detection apparatus according to claim 1 , wherein the circuitry is further configured to:
detect the first motion vector by calculation of a candidate motion vector using a projection matrix corresponding to an angle of view and an aspect ratio each corresponding to the depth image data and the captured image data, and
merge the candidate motion vector with a previously detected motion vector.
3. A motion vector detection method, comprising:
acquiring, by a motion sensor, self-motion data, depth image data, and captured image data;
detecting, by a circuitry, a first motion vector based on the self-motion data and the depth image data;
detecting, by the circuitry, a second motion vector based on the captured image data;
calculating, by the circuitry, a first predicted error of the first motion vector and a second predicted error of the second motion vector;
calculating, by the circuitry, a first degree of reliability and a second degree of reliability based on the first predicted error and the second predicted error, respectively; and
merging, by the circuitry, the first motion vector with the second motion vector based on the first degree of reliability and the second degree of reliability.