Image-processing device for estimating the motion of objects situated in said image

• US 4,760,445 A
• Filed: 04/09/1987
• Issued: 07/26/1988
• Est. Priority Date: 04/15/1986
• Status: Expired due to Fees

• Current Assignee: US Philips Corporation (Koninklijke Philips N.V.)
• Original Assignee: US Philips Corporation (Koninklijke Philips N.V.)
• Inventors: Mijiyawa, Mekano
• Primary Examiner(s): Britton, Howard W.
• Application Number:
  US07/036,221
• Time in Patent Office: 474 Days
• Field of Search: 358/135, 358/136, 358/133, 358/138, 358/105, 375/27
• US Class Current: 348/699
• CPC Class Codes: G06T 7/277 involving stochastic approaches, e.g. using Kalman filters
• CPC Class Codes: H04N 19/577 Motion compensation with bidirectional frame interpolation, i.e. using B-pictures

Abstract

A TV image processing device includes a Kalman filtering estimation member (50) which provides for each image point "i" having coordinates Z_i a displacement vector DK(Z_i) based on information contained in two successive images I(Z_i,t) and I(Z_i,t-TR) separated by a period TR. It further includes a second estimation member (60) for providing for each point "i" a displacement vector DB(Z_i), a quality measuring member (70) for evaluating the vectors DK(Z_i) and DB(Z_i) and for influencing a selector (80) for the vectors so as to select that one of the two vectors DB(Z_i) and DK(Z_i) which provides the best estimation quality for the output vector D(Z_i). The Kalman filtering member has an input (90) for receiving the selected vector D(Z_i).

Claims

1. An image-processing device for estimating the motion displacement D(Z_i) of objects situated in the image comprising a first estimation member which is an estimation member of a KALMAN filtering member providing for each image point "i" having the coordinates Z_i, a displacement vector DK(Z_i) from two successive images I(Z_i, t) and I(Z_i, t-TR) separated by a period TR, a second estimation member for providing a displacement vector DB(Z_i) for each point "i" of the image, a quality measuring member for evaluating the vectors DK(Z_i) and DB(Z_i) and for controlling a vector selector to select that one of the two vectors DB(Z_i) and DK(Z_i) which provides the best estimation for the vector D(Z_i), such that the first estimation member is adapted to receive said vector D(Z_i), wherein means forming part of the Kalman filtering member provide the value DK(Z_i) from predetermined parameters and means providing a value referred to as the innovation IFD so that

View Dependent Claims (2, 3)

2. An image-processing device for estimating the displacement D(Z_i) of objects situated in said image as claimed in claim 1 further comprising a device for providing a value IFDB(Z_i)=I(Z_i, t)-I(Z_i -DB(Z_i), t-TR) such that this value is utilized as a quality criterion by the quality measuring member, whereby D(Z_i) is taken to be the value of DB(Z_i) or DK(Z_i) which gives the smallest value of IFD(Z_i) or IFDB(Z_i).

3. An image-processing device for estimating the displacement D(Z_i) of objects situated in said image as claimed in claim 1, further comprising a motion detector for providing when there is no motion of objects in the image, a vector D(Z_i) having a value indicating that there is no motion.

4. An image-processing device for estimating the motion displacement D(Z_i) of objects situated in the image comprising a first estimation member providing for each image point "i" having the coordinates Z_i, a displacement vector DK(Z_i) from two successive images I(Z_i, t) and I(Z_i, t-TR) separated by a period TR, a second estimation member operating by means of block correlation for providing a displacement vector DB(Z_i) for each point "i" of the image, a quality measuring member for evaluating the vectors DK(Z_i) and DB(Z_i) for controlling a vector selector to select that one of the two vectors DB(Z_i) and DK(Z_i) which provides the best estimation for the vector D(Z_i), such that the first estimation member is adapted to receive said vector D(Z_i), wherein means forming part of a Kalman filtering member provide the value DK(Z_i) from predetermined parameters and means providing a value referred to as the innovation IFD so that

View Dependent Claims (5, 6)

5. An image-processing device for estimating the displacement D(Z_i) of objects situated in said image as claimed in claim 4, further comprising a motion detector for providing, when there is no motion of objects in the image, a vector D(Z_i) having a value indicating that there is no motion.

6. An image-processing device for estimating the displacement D(Z_i) of objects situated in said image as claimed in claim 4, further comprising a device for providing a value IFDB(Z_i)=I(Z_i, t)-I(Z_i -DB(Z_i), t-TR) such that this value is utilized as a quality criterion by the quality measuring member, whereby D(Z_i) is taken to be the value of DB(Z_i) or DK(Z_i) which gives the smallest value of IFD(Z_i) or IFDB(Z_i).

7. An image-processing device for estimating the displacement D(Z_i) of objects situated in said image comprising a first estimation member providing for each image point "i" having the coordinates Z_i, a displacement vector DK(Z_i) from two successive images I(Z_i, t) and I(Z_i, t-TR) separated by a period TR, a second estimation member for providing a displacement vector DB(Z_i) for each point "i" of the image, a quality mesuring member for evaluating the vectors DK(Z_i) and DB(Z_i) and for controlling a vector selector to select that one of the two vectors DB(Z_i) and DK(Z_i) which provides the best estimation for the vector D(Z_i), such that the first estimation member is adapted to receive said vector D(Z_i), further comprising a motion detector for providing when there is no motion of objects in the image, a vector D(Z_i) having a value indicating that there is no motion, and further comprising a device for providing a value IFDB(Z_i)=I(Z_i, t)-I(Z_i -DB(Z_i), t-TR) such that this value is utilized as a quality criterion by the quality measuring member, whereby D(Z_i) is taken to be the value of DB(Z_i) or DK(Z_i) which gives the smallest value of IFD(Z_i) or IFDB(Z_i).

8. An image-processing device for estimating the displacement D(Z_i) of objects situated in said image comrising a first estimation member which is an estimation member of a KALMAN filtering member providing for each image point "i" having the coordinates Z_i, a displacement vector DK(Z_i) from two successive images I(Z_i, t) and I(Z_i, t-TR) separated by a period TR, a second estimation member for providing a displacement vector DB(Z_i) for each point "i" of the image, a quality measuring member for evaluating the vectors DK(Z_i) and DB(Z_i) and for controlling a vector selector to select that one of the two vectors DB(Z_i) and DK(Z_i) which provides the best estimation for the vector D(Z_i), such that the first estimation member is adapted to receive said vector D(Z_i), further comprising a device for providing a value IFDB(Z_i)=I(Z_i, t)-I(Z_i -DB(Z_i), t-TR) such that this value is utilized as a quality criterion by the quality measuring member, whereby D(Z_i) is taken to be the value of DB(Z_i) or DK(Z.sub. i) which gives the smallest value of IFD(Z_i) or IFDB(Z_i).

9. An image-processing device for estimating the displacement D(Z_i) of objects situated in said image comprising a first estimation member providing for each image point "i" having the coordinates Z_i, a displacement vector DK(Z_i) from two successive images I(Z_i, t) and I(Z_i, t-TR) separated by a period TR, a second estimation member operating by means of block correlation for providing a displacement vector DB(Z_i) for each point "i" of the image, a quality measuring member for evaluating the vectors DK(Z_i) and DB(Z_i) and for controlling a vector selector to select that one of the two vectors DB(Z_i) and DK(Z_i) which provides the best estimation for the vector D(Z_i), such that the first estimation member is adapted to receive said vector D(Z_i), further comprising a device for providing a value IFDB(Z_i)=I(Z_i, t)-I(Z_i -DB(Z_i), t-TR) such that this value is utilized as a quality criterion by the quality measuring member, whereby D(Z_i) is taken to be the value of DB(Z_i) or DK(Z_i) which gives the smallest value of IFD(Z_i) or IFDB(Z_i).