Systems and methods for luminance-based scene-change detection for continuous autofocus
View Patent ↗Imaging systems with image sensors and image processing circuitry are provided. The image processing circuitry may identify motion and perform autofocus (e.g., continuous autofocus) using images captured by an image sensor. Auto exposure metrics such as average luminance values and autofocus statistics such as sharpness scores may be calculated for each image. The auto exposure metrics may be used to calculate motion scores and identify directional motion between a series of captured images. The motion scores may be used with the sharpness scores to determine when to perform autofocus functions such as when to refocus a lens for a continuous autofocus application. For example, the motion scores may be monitored to identify motion that exceeds a given magnitude and duration. After identification of motion, motion scores and sharpness scores may be used to determine when a given scene has stabilized and when the lens should be refocused.
1. A method, comprising:
focusing incident light on a pixel array with a lens;
capturing a sequence of images with the pixel array;
with image processing circuitry, calculating statistical metrics from the captured sequence of images;
calculating a motion score based on the statistical metrics; and
determining when to refocus the lens based at least partly on the motion score, wherein determining when to refocus the lens based at least partly on the motion score comprises:
incrementing a motion count that is indicative of how many images from the sequence of images have a motion score that is greater than a first threshold;
in response to identifying that the motion count is greater than a second, motion counter threshold, incrementing a stable count that is indicative of how many images from the sequence of images have a motion score that is less than the first threshold.
2. The method defined in claim 1 , wherein determining when to refocus the lens further comprises identifying a reference image from the sequence of images.
3. The method defined in claim 2 , wherein determining when to refocus the lens further comprises identifying that the stable count is greater than a third threshold.
4. The method defined in claim 3 , wherein determining when to refocus the lens further comprises:
calculating a focus-change value based on a focus score of the reference image and a focus score of a given image from the sequence of images; and
identifying that the focus-change value is greater than a fourth threshold, the method further comprising:
in response to identifying that the focus-change value is greater than the fourth threshold, refocusing the lens.
5. A method for identifying motion from a first image and a second image using image processing circuitry, comprising:
calculating a first set of values associated with automatic exposure metrics of the first image;
calculating a second set of values associated with automatic exposure metrics of the second image;
calculating a first set of motion scores based on differences between the values in the first set of values and the values in the second set of values, wherein each motion score of the first set of motion scores is associated with a direction of motion;
calculating a third set of values associated with automatic exposure metrics of a third image;
calculating a second set of motion scores based on differences between the values in the second set of values and the values in the third set of values, wherein each motion score of the second set of motion scores is associated with a direction of motion;
identifying motion in a first direction based on the first set of motion scores;
identifying motion in a second direction based on the second set of motion scores;
identifying that the first direction is equal to the second direction; and
in response to identifying that the first direction is equal to the second direction, refocusing a lens.
6. The method defined in claim 5 wherein calculating the first set of values comprises calculating zone luminance values.
7. The method defined in claim 5 wherein calculating the first set of motion scores comprises:
forming a first subset from the first set of values and a second subset from the second set of values; and
calculating a motion score based on a difference between a sum of differences between the first subset and the second subset.
8. The method defined in claim 5 further comprising:
identifying a highest motion score from the first set of motion scores;
identifying a second-highest motion score from the first set of motion scores; and
calculating a final motion score based on the highest motion score and the second-highest motion score.
9. The method defined in claim 8 further comprising:
identifying that the final motion score exceeds a motion threshold.
10. The method defined in claim 9 further comprising:
identifying a direction of motion associated with the final motion score.
11. The method defined in claim 5 further comprising:
calculating a first sharpness score from the first image;
calculating a second sharpness score from the second image; and
calculating a difference between the first and second sharpness scores, wherein refocusing the lens comprises refocusing the lens in response to identifying that the difference is greater than a threshold value.