Aligning Two Images By Matching Their Feature Points

A method of digital image signal processing, comprising:

• (a) providing a multiple camera imaging system including a first camera sensor, a second camera sensor, and an image signal processor (ISP);
• (b) using the first camera sensor to obtain a first image of a scene from a first vantage perspective point;
• using the second camera sensor to obtain a second image of the scene from a second vantage perspective point;
• (c) using the ISP to produce a first roughly-aligned image from the first image, and a second roughly-aligned image from the second image;
• (d) using the ISP to utilize the first and second roughly-aligned images to produce a disparity image;
• (e) using the ISP to identify a feature point within the first roughly-aligned image;
• to utilize the disparity image to create a search zone within the second roughly-aligned image;
• and to identify a group of candidate feature points within the search zone;
• (f) using the ISP to identify within the search zone a best-matched candidate feature point that best matches the feature point within the first roughly-aligned image to form a best-matched feature point pair;
• and
• (g) using the ISP to utilize information from the best-matched feature point pair to further align the first and second roughly-aligned images.

A multiple camera imaging system, comprising:

• a first camera image sensor configured to obtain a first image of a scene from a first vantage perspective point;
• a second camera image sensor configured to obtain a second image of the scene from a second vantage perspective point;
• and
• an image signal processor (ISP), including:
• (a) a preliminary image processing block configured to produce a first roughly-aligned image from the first image, and a second roughly-aligned image from the second image;
• (b) a disparity image creation block configured to produce a disparity image by using the first and second roughly-aligned images;
• (c) a first feature point selector block configured to identify a feature point within the first roughly-aligned image;
• (d) a disparity value generator block configured to produce a disparity value by using the feature point and the disparity image;
• (e) a search zone generator block configured to produce, within the second roughly-aligned image, a search zone by using the feature point and the disparity value;
• (f) a second feature point selector block configured to identify a group of candidate feature points within the search zone;
• (g) a feature point matching block configured to identify, within the search zone, a best-matched candidate feature point that best matches the feature point within the first roughly-aligned image to form a best-matched feature point pair;
• wherein the ISP is further configured to use information from the best-matched feature point pair to further align the first and second roughly-aligned images.