# US 2010/0303340 A1 · Stereo-image registration and change detection system and method > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2010/0303340 A1 - Document ID: US20100303340A1 - Status: Published application - Kind: A1 - Application number: 12/739,592 - Filed: 2008-10-23 - Published: 2010-12-02 - Earliest claimed: 2007-10-23 - Assignees: Elta Systems Ltd. - Inventors: Michael Abraham; Victor Gostynski - Web page: https://patentgrep.com/patent/US20100303340A1 ## CPC - G06T 7/285: using a sequence of stereo image pairs - G06T 7/593: from stereo images - G06T 2207/10021: Stereoscopic video; Stereoscopic image sequence - G06T 2207/30181: Earth observation - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL ## Abstract A system and method for registering stereoscopic images comprising: obtaining at least two sets of stereoscopic images, each one of the at least two sets including at least two images that are taken from different angles, determining at least two groups of images, each one of the groups including at least two images that are respective images of at least two of the sets or are derived therefrom. For each one of the groups, calculating a respective optimal entities list and stereo-matching at least two images, each one being or derived from different one of the at least two groups and same or different sets, using at least four optimal entities from each one of the optimal entities list, thereby giving rise to at least one pair of registered stereoscopic images. ## Claims ### Claim 1 A method for registering stereoscopic images, the method comprising: a) obtaining at least two sets of stereoscopic images, each one of said at least two sets including at least two images that are taken from different angles; b) determining at least two groups of images, each one of said groups including at least two images that are respective images of at least two of said sets or are derived therefrom; and c) for each one of said groups, calculating a respective optimal entities list; d) stereo-matching at least two images, each one being or derived from different one of said at least two groups and same or different sets, using at least four optimal entities from each one of said optimal entities list, thereby giving rise to at least one pair of registered stereoscopic images. ### Claim 2 (depends on 1) The method according to claim 1 wherein said determining at least two groups of images further comprises: a) associating at least two images from said at least two sets with a first collection of images; b) iteratively, until a condition is met: i. selecting one image from said first collection and transferring it to a new group giving rise to a reference image; ii. conducting a series of registration attempts, for each one of said registration attempts attempting to register the reference image with each image from said first collection; iii. computing a normalized similarity measure for each one of the registration attempts; iv. computing a similarity threshold; and v. transferring each image from said first collection having said normalized similarity measure above said similarity threshold to the new group. ### Claim 3 (depends on 2) The method according to claim 2 wherein the condition is met when there are no images associated with the first collection. ### Claim 4 (depends on 1) The method according to claim 1 wherein said stereo-matching further comprising: a) computing a normalized similarity measure for each one of said pairs of registered stereoscopic images; b) providing a similarity threshold; wherein successful stereo-match of each of the pairs of registered stereoscopic images is confirmed if the corresponding normalized similarity measure exceeds the corresponding similarity threshold. ### Claim 5 (depends on 1) The method according to claim 1, wherein each one of said at least two sets including at least two images that are taken during the same flight leg. ### Claim 6 (depends on 1) The method according to claim 1, wherein each of said groups include at least two images that (i) are respective images of at least two of said sets or are derived therefrom, and (ii) are taken from acquisition angles within a range of 10 degrees or less. ### Claim 7 (depends on 1) The method according to claim 1, wherein each one of said at least two sets including at least two images that are taken at substantially the same time. ### Claim 8 (depends on 1) The method according to claim 1, wherein said calculating a respective optimal entities list, stipulated in (c) comprises: a) registering at least two of said group's images; b) detecting changes between the registered images thereby giving rise to a candidate entities list; and c) selecting at least four optimal entities from said candidate entities list in accordance with at least one selection criterion, thereby giving rise to said optimal entities list. ### Claim 9 (depends on 1) The method according to claim 1, further comprising: a) calculating seed correspondence points for each of the optimal entities from the optimal entities list; wherein four or more of said seed correspondence points calculated from each one of said optimal entities list are configured to be fed to a stereo-matching module which stereo-matches at least two images, each one being or derived from different one of said at least two groups and same or different sets, using said seed correspondence points, thereby giving rise to at least one pair of registered stereoscopic images. ### Claim 10 (depends on 1) The method according to claim 1 wherein the pair of registered stereoscopic images is produced from stereoscopic images that are being or derived from images taken from two angles, at least 10 degrees apart. ### Claim 11 (depends on 1) The method according to claim 1 wherein the pair of registered stereoscopic images is produced from stereoscopic images that are being or derived from images taken from two angles, at least 25 degrees apart. ### Claim 12 (depends on 1) The method according to claim 1 wherein the pair of registered stereoscopic images is produced from stereoscopic images that are being or derived from images taken from two angles, at least 35 degrees apart. ### Claim 13 (depends on 1) The method according to claim 1 wherein the pair of registered stereoscopic images is produced from stereoscopic images that are being or derived from images taken from two angles, at least 45 degrees apart. ### Claim 14 A method for detecting differences in a photographed scenery between two time points, the method comprising: a) obtaining at least two sets of stereoscopic images, each one of said at least two sets including at least two images that are taken from different angles; b) determining at least two groups of images, each group including at least two images that are respective images of at least two of said sets or are derived therefrom; and c) for each one of said groups, calculating a respective optimal entities list; d) stereo-matching at least two images, each one being or derived from different one of said at least two groups and same or different sets, using at least four optimal entities from each one of said optimal entities list giving rise to at least one pair of registered stereoscopic images; e) producing depth maps for at least two of said pairs of registered stereoscopic images; and f) calculating height differences between corresponding regions of said depth maps; whereby volumetric differences are detected between said sets of stereoscopic images. ### Claim 15 (depends on 14) The method according to claim 14, wherein each of said group includes at least two images that (i) are respective images of at least two of said sets or are derived therefrom, and (ii) are taken from acquisition angles within a range of 10 degrees or less. ### Claim 16 A method for detecting differences in a photographed scenery between two time points, the method comprising: a) obtaining at least two sets of stereoscopic images, each one of said at least two sets including at least two images that are taken from different angles; b) determining at least two groups of images, each group including at least two images that are respective images of at least two of said sets or are derived therefrom; c) for each one of said groups, calculating a respective optimal entities list; d) stereo-matching at least two images, each one being or derived from different one of said at least two groups and same or different sets, using at least four optimal entities from each at least four optimal entities from each one of said optimal entities list of said optimal entities list giving rise to at least one pair of registered stereoscopic images; e) producing depth maps for at least two of said pairs of registered stereoscopic images, at least within regions derived from coordinates of candidate entities of said candidate entities list; and f) reassessing candidate entities from said candidate entities list against at least two of said depth maps; and wherein said candidate entities are classified as volumetric differences if the reassessment shows height difference. ### Claim 17 (depends on 16) The method according to claim 16, wherein each of said group includes at least two images that (i) are respective images of at least two of said sets or are derived therefrom, and (ii) are taken from acquisition angles within a range of 10 degrees or less. ### Claim 18 A method for detecting differences in a photographed scenery between two time points, the method comprising: a) obtaining at least two sets of stereoscopic images of said scenery, each one of said at least two sets including at least two images that are taken from different angles; b) generating at least two pairs of registered stereoscopic images, each one of said pairs includes stereoscopic images from one of said sets respectively; c) calculating a depth map for each one of said pairs of registered stereoscopic images; and d) calculating height differences between corresponding regions of said depth maps; whereby volumetric differences are detected between said sets of stereoscopic images. ### Claim 19 (depends on 18) The method according to claim 18 wherein each of said pairs of registered stereoscopic images are produced from stereoscopic images that are being or derived from images taken from two angles, at least 10 degrees apart. ### Claim 20 (depends on 18) The method according to claim 18 wherein each of said pairs of registered stereoscopic images are produced from stereoscopic images that are being or derived from images taken from two angles, at least 25 degrees apart. ### Claim 21 (depends on 18) The method according to claim 18 wherein each of said pairs of registered stereoscopic images are produced from stereoscopic images that are being or derived from images taken from two angles, at least 35 degrees apart. ### Claim 22 (depends on 18) The method according to claim 18 wherein each of said pairs of registered stereoscopic images are produced from stereoscopic images that are being or derived from images taken from two angles, at least 45 degrees apart. ### Claim 23 A system for registering stereoscopic images, the system comprising: a processor and associated storage, said processor and associated storage are configured to perform at least the following: a) obtain at least two sets of stereoscopic images, each one of said at least two sets including at least two images that are taken from different angles; b) determine at least two groups of images, each group including at least two images that are respective images of at least two of said sets or are derived therefrom; and c) for each one of said groups, calculate a respective optimal entities list; wherein at least four optimal entities from each one of said optimal entities list are configured to be fed to a stereo-matching module which stereo-matches at least two images, each one being or derived from different one of said at least two groups and same or different sets, using said optimal entities, thereby giving rise to at least one pair of registered stereoscopic images. ### Claim 24 (depends on 23) The system according to claim 23, wherein each of said group includes at least two images that (i) are respective images of at least two of said sets or are derived therefrom, and (ii) are taken from acquisition angles within a range of 10 degrees or less. ### Claim 25 A system for detecting differences in a photographed scenery between two time points, the system comprising: a processor and associated storage, said processor and associated storage are configured to perform at least the following: a) obtain at least two sets of stereoscopic images of said scenery, each one of said at least two sets including at least two images that are taken from different angles; b) generate at least two pairs of registered stereoscopic images, each one of said pairs includes stereoscopic images from one of said sets respectively; c) calculate a depth map for each one of said pairs of registered stereoscopic images; and d) calculate height differences between corresponding regions of said depth maps; whereby volumetric differences are detected between said sets of stereoscopic images. ### Claim 26 A program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform method operations for registering stereoscopic images, comprising: a) obtaining at least two sets of stereoscopic images, each one of said at least two sets including at least two images that are taken from different angles; b) determining at least two groups of images, each group including at least two images that are respective images of at least two of said sets or are derived therefrom; and c) for each one of said groups, calculating a respective optimal entities list; wherein at least four optimal entities from each one of said optimal entities list are configured to be fed to a stereo-matching module which stereo-matches at least two images, each one being or derived from different one of said at least two groups and same or different sets, using said optimal entities, thereby giving rise to at least one pair of registered stereoscopic images. ### Claim 27 (depends on 24) The program storage device according to claim 24, wherein each of said group includes at least two images that (i) are respective images of at least two of said sets or are derived therefrom, and (ii) are taken from acquisition angles within a range of 10 degrees or less. ### Claim 28 A program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform method operations for detecting differences in a photographed scenery between two time points comprising: a) obtaining at least two sets of stereoscopic images of said scenery, each one of said at least two sets including at least two images that are taken from different angles and; b) generating at least two pairs of registered stereoscopic images, each one of said pairs includes stereoscopic images from one of said sets respectively; c) calculating a depth map for each one of said pairs of registered stereoscopic images; and d) calculating height differences between corresponding regions of said depth maps; whereby volumetric differences are detected between said sets of stereoscopic images. ### Claim 29 A method for registering stereoscopic images, the method comprising: a) obtaining at least two sets of stereoscopic images, each one of said at least two sets including at least two images that are taken from different angles; b) determining at least two groups of images, each group including at least two images that are respective images of at least two of said sets or are derived therefrom; and c) for each one of said groups: i) registering at least two of said group's images; ii) detecting changes between the registered images thereby giving rise to a candidate entities list; and iii) selecting at least four optimal entities from said candidate entities list in accordance with at least one selection criterion, thereby giving rise to said optimal entities list; wherein at least four optimal entities from each one of said optimal entities list are configured to be fed to a stereo-matching module which uses said optimal entities to stereo-match at least two images, each one being or derived from different one of said at least two groups, thereby a successful stereo-match of said at least two images is achieved. ### Claim 30 (depends on 27) The method according to claim 27, further comprising: a) producing depth maps for at least two of said pairs of registered stereoscopic images; and b) calculating height differences between corresponding regions of said depth maps whereby volumetric differences are detected between said sets of stereoscopic images.