# US 10,319,108 B2 · System and method for machine vision object orientation measurement > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 10,319,108 B2 - Document ID: US10319108B2 - Status: Granted - Kind: B2 - Application number: 15/432,909 - Filed: 2017-02-14 - Published: 2018-08-16 - Granted: 2019-06-11 - Earliest claimed: 2017-02-14 - Assignees: JX Imaging Arts, LLC - Applicants: JX Imaging ARTS, LLC - Inventors: Jan Segert - Web page: https://patentgrep.com/patent/US10319108B2 ## CPC - G06T 7/70: Determining position or orientation of objects or cameras (camera calibration G06T7/80) - G06T 7/11: Region-based segmentation - G06T 7/13: Edge detection - G06T 7/75: {involving models} - G06T 7/77: using statistical methods - G06T 2207/10004: Still image; Photographic image - G06T 2207/10016: Video; Image sequence - G06T 2207/10024: Color image - G06T 2207/10116: X-ray image - G06T 2207/30052: Implant; Prosthesis - G06T 2207/30221: Sports video; Sports image - G06T 2207/30224: Ball; Puck - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL ## Abstract Disclosed are methods and systems for measuring orientation parameters of spherical objects by image processing. Geometric shapes detected in a captured image are measured by generating and scoring models of the geometric shapes, where a high numerical score indicates a good match of the model to the image. The orientation parameters are computed from the geometric parameters measured from the image. A fast, precise, and accurate measurement requires generation and scoring be computationally efficient, meaning that many high-scoring models are generated in a short period of time. Disclosed are embodiments for a multistep measurement process for measuring the orientation parameters of an artificial hip joint from an x-ray image, measuring rotation of a basketball from video frame images, and navigating a spacecraft utilizing sunlight incident on a planetary body. ## Claims ### Claim 1 A method for measuring orientation parameters of a physical object comprising: acquiring a first image of the physical object, wherein at least a portion of the physical object is spherical; selecting a first region of interest in the first image; performing a circular random sample consensus operation on the first region of interest to detect a reference circle; cropping the first region of interest to the image portion inside the reference circle; performing a first semiellipse random sample consensus operation on the cropped first region of interest to detect a first primary semiellipse; computing a first set of measured orientation parameters of the object from geometric parameters of the first primary semiellipse, wherein the first primary semiellipse indicates at least a portion of a first primary great circle discernible on the physical object; whereby the orientation parameters of the physical object are measured from the acquired image. ### Claim 2 (depends on 1) A method as in claim 1 wherein the first circular random sample consensus operation comprises an initial first circular random sample consensus operation establishing an initial reference circle and a final first circular random sample consensus operation which is performed on the first region of the image cropped to a first annular ring comprising image points near the initial reference circle wherein the final first circular consensus operation establishes the reference circle. ### Claim 3 (depends on 1) A method as in claim 1 wherein the first semiellipse random sample consensus operation comprises an initial first semiellipse random sample consensus operation establishing an initial first semiellipse and a final first semiellipse random sample consensus operation which is performed on a region of the image cropped to a semiellipse region comprising image points near the initial first semiellipse wherein the final first semiellipse consensus establishes the first semiellipse. ### Claim 4 (depends on 1) A method as in claim 1 wherein the first semiellipse random sample consensus operation is constrained by the reference circle, wherein the resulting first constrained semiellipse random sample consensus operation is performed by selecting no more than 2 random edge points. ### Claim 5 (depends on 1) A method as in claim 1 wherein the image is an x-ray. ### Claim 6 (depends on 1) A method as in claim 1 wherein the physical object is hemispherical, and wherein the first primary great circle discernible on the object is an edge of the hemispherical object. ### Claim 7 (depends on 1) A method as in claim 1 wherein the physical object is hemispherical and wherein the set of measured orientation parameters comprise at least version and inclination angles. ### Claim 8 (depends on 1) A method as in claim 1 wherein the physical object is an acetabular component of an artificial hip joint. ### Claim 9 (depends on 1) A method as in claim 1 wherein the physical object is a basketball. ### Claim 10 (depends on 1) A method as in claim 1 wherein the orientation parameters are angles relative to a fixed frame of reference. ### Claim 11 (depends on 1) A method as in claim 1 further comprising: performing a second semiellipse random sample consensus operation on the cropped first region of interest which has been further cropped to remove image points near the detected first primary semiellipse to detect a first secondary semiellipse; computing a second set of measured orientation parameters of the physical object from geometric parameters of the first secondary semiellipse; whereby the first set of measured orientation parameters and the second set of measured orientation parameters of the physical object are measured orientation parameters of the first primary great circle and a first secondary great circle discernible on the physical object from the first image. ### Claim 12 (depends on 11) A method as in claim 11 further comprising: acquiring at least one image after the first image forming a first series of images, wherein the images in the first series of images are sequential in time; repeating the method of claim 11 on each subsequent acquired image in the first series of images; calculating a first set of measured rotational parameters for the first series of images. ### Claim 13 (depends on 11) A method as in claim 11 further comprising: acquiring at least one image after the first image forming a first series of images, wherein the images in the first series of images are sequential in time; repeating the method of claim 11 on each subsequent acquired image in the first series of images; calculating a first set of measured rotational parameters for the first series of images; acquiring at least one image after the first series of images forming a second series of images, wherein the images in the second series of images are sequential in time; repeating the method of claim 11 on each subsequent acquired image in the second series of images; calculating a second set of measured rotational parameters for the second series of images; comparing at least one measured rotational parameter from the first set of measured rotational parameters to at least one measured rotational parameter from the second set of measured rotational parameters. ### Claim 14 (depends on 13) A method as in claim 13 wherein the first series of images and the second series of images each comprise 2 images; whereby the first and second sets of measured rotational parameters may be reliably calculated when the object rotation angle is less than a threshold condition. ### Claim 15 (depends on 13) A method as in claim 13 wherein the first series of images and the second series of images each comprise at least 3 images; whereby the first and second sets of measured rotational parameters may be reliably calculated for any rotational conditions. ### Claim 16 (depends on 11) A method as in claim 11 wherein all the performed semiellipse random sample consensus operations are constrained by the respective circles and wherein all of the constrained random sample consensus operations are performed by selecting no more than 2 edge points within the region for each operation performed. ### Claim 17 (depends on 1) A method as in claim 1 further comprising: performing a second semiellipse random sample consensus operation on the cropped first region of interest to detect a first secondary semiellipse, wherein at least one measured orientation parameter of the first primary great circle discernible on the object and associated with the first primary semiellipse is utilized to constrain the second semiellipse random sample consensus operation, wherein with prior knowledge that a first secondary great circle is orthogonal to the first primary great circle, the constrained second semiellipse random sample consensus operation utilizes no more than a single randomly selected edge point computing a second set of measured orientation parameters of the object from geometric parameters of the first secondary semiellipse; whereby the first set of measured orientation parameters and the second set of measured orientation parameters of the object are measured orientation parameters of the first primary great circle and a first secondary great circle discernible on the object from the first image. ### Claim 18 A system for measuring the orientation of an object comprising: an image acquisition component wherein an image of the object is acquired, wherein at least a portion of the object is spherical; a central processing unit configured to execute instructions comprising: an image processing selection module configured to select a region of interest in the image, an image processing object identification module configured to perform a circular random sample consensus operation on the region of interest to detect a reference circle, to crop the region of interest to the image portion inside the reference circle, to perform a semiellipse random sample consensus operation on the cropped region of interest to detect a semiellipse, wherein the semiellipse indicates at least a portion of a first primary great circle discernible on the object, to compute a set of measured orientation parameters of the object from geometric parameters of the semiellipse. ### Claim 19 (depends on 18) A system as in claim 18 wherein the circular random sample consensus operation comprises an initial circular random sample consensus operation establishing an initial reference circle and a final circular random sample consensus operation which is performed on a region of the image cropped to an annular ring comprising image points near the initial reference circle wherein the final circular consensus operation establishes the reference circle. ### Claim 20 (depends on 18) A system as in claim 18 wherein the semiellipse random sample consensus operation comprises an initial semiellipse random sample consensus operation establishing an initial semiellipse and a final semiellipse random sample consensus operation which is performed on a region of the image cropped to images points near the initial semiellipse, wherein the final semiellipse consensus operation establishes the semiellipse. ### Claim 21 (depends on 18) A system as in claim 18 wherein the semiellipse random sample consensus operation is constrained by the circle, wherein the constrained semiellipse random sample consensus operation is performed by selecting no more than 2 edge points within the region. ### Claim 22 (depends on 18) A system as in claim 18 wherein the image is an x-ray. ### Claim 23 (depends on 18) A system as in claim 18 wherein the object is an acetabular component of an artificial hip joint.