Contour detecting apparatus and method, and storage medium storing contour detecting program
A contour detecting apparatus for accurately detecting a contour of a target object without increasing the detection time is disclosed. The apparatus comprises a section for storing a contour model consisting of nodes, which surrounds at least one target object in a captured image and is used for detecting a contour of the target object; a deforming section for contracting or expanding the contour model by shifting the nodes based on a predetermined rule; a section for calculating a distance between two non-adjacent nodes of the contour model which was deformed by the deforming section, and determining that the contour model is to be split when the calculated distance is equal to or smaller than a predetermined threshold; and a section for splitting the contour model according to a result of the above determination, wherein the splitting operation is executed in the vicinity of said non-adjacent nodes.
1. A contour detecting apparatus comprising:
a contour model storage section for storing a contour model consisting of a plurality of nodes, which surrounds at least one target object included in a captured image and is used for detecting a contour of said target object;
a contour model deforming section for contracting or expanding the contour model by shifting the nodes based on a predetermined rule for deformation;
a split-determining section for calculating a distance between two non-adjacent nodes among the nodes of the contour model which was deformed by the contour model deforming section, and determining that the contour model is to be split when the calculated distance is equal to or smaller than a predetermined threshold value; and
a contour model splitting section for splitting the contour model according to a result of the determination executed by the split-determining section, wherein the splitting operation is executed in the vicinity of said two non-adjacent nodes.
2. The contour detecting apparatus as claimed in claim 1 , further comprising:
a distance image obtaining section for obtaining a distance image corresponding to said captured image by acquiring distance data of each pixel in said captured image, and wherein:
the contour model deforming section deforms the contour model based on the obtained distance image, so as to make the contour model correspond only to said target object which is present in a predetermined distance range; and
the split-determining section determines whether the contour model defined in the predetermined distance range is to be split.
3. The contour detecting apparatus as claimed in claim 1 , wherein the split-determining section calculates the square of the distance between said two non-adjacent nodes and determines that the contour model is to be split when the calculated square of the distance is equal to or smaller than the predetermined threshold value.
4. The contour detecting apparatus as claimed in claim 1 , wherein the contour model deforming section deforms the contour model in a manner such that an energy function assigned to the contour model has a minimum solution.
5. The contour detecting apparatus as claimed in claim 1 , wherein when a thinnest portion of said target object is known, the threshold value is predetermined in consideration of the thickness of the thinnest portion, so as to prevent the contour model from being split at the thinnest portion.
6. A contour detecting apparatus, comprising:
a contour model storage section for storing a contour model consisting of a plurality of nodes, which surrounds at least one target object included in a captured image and is used for detecting a contour of said target object;
a contour model deforming section for contracting or expanding the contour model by shifting the nodes based on a predetermined rule for deformation;
a split-determining section for calculating a distance between two non-adjacent nodes among the nodes of the contour model which was deformed by the contour model deforming section, and determining that the contour model is to be split when the calculated distance is equal to or smaller than a predetermined threshold value; and
a distance image obtaining section for obtaining a distance image corresponding to said captured image by acquiring distance data of each pixel in said captured image, wherein:
the contour model deforming section deforms the contour model based on the obtained distance image, so that the contour of said target object is detected.
7. A contour detecting method comprising the steps of:
a contour model deforming step of contracting or expanding a contour model consisting of a plurality of nodes, which surrounds at least one target object included in a captured image and is used for detecting a contour of said target object, by shifting the nodes based on a predetermined rule for deformation;
a split-determining step of calculating a distance between two non-adjacent nodes among the nodes of the contour model which was deformed in the contour model deforming step, and determining that the contour model is to be split when the calculated distance is equal to or smaller than a predetermined threshold value; and
a contour model splitting step of splitting the contour model according to a result of the determination executed in the split-determining step, wherein the splitting operation is executed in the vicinity of said two non-adjacent nodes.
8. The contour detecting method as claimed in claim 7 , further comprising:
a distance image obtaining step of obtaining a distance image corresponding to said captured image by acquiring distance data of each pixel in said captured image, and wherein:
the contour model deforming step includes deforming the contour model based on the obtained distance image, so as to make the contour model correspond only to said target object which is present in a predetermined distance range; and
the split-determining step includes determining whether the contour model defined in the predetermined distance range is to be split.
9. The contour detecting method as claimed in claim 7 , wherein the split-determining step includes calculating the square of the distance between said two non-adjacent nodes and determining that the contour model is to be split when the calculated square of the distance is equal to or smaller than the predetermined threshold value.
10. The contour detecting method as claimed in claim 7 , wherein the contour model deforming step includes deforming the contour model in a manner such that an energy function assigned to the contour model has a minimum solution.
11. The contour detecting method as claimed in claim 7 , wherein when a thinnest portion of said target object is known, the threshold value is predetermined in consideration of the thickness of the thinnest portion, so as to prevent the contour model from being split at the thinnest portion.
12. A contour detecting method, comprising the steps of:
a contour model deforming step of contracting or expanding a contour model consisting of a plurality of nodes, which surrounds at least one target object included in a captured image and is used for detecting a contour of said target object, by shifting the nodes based on a predetermined rule for deformation;
a calculating step for calculating a distance between two non-adjacent nodes among the nodes of the contour model;
a determining step for determining that the contour model is to be split when the calculated distance is equal to or smaller than a predetermined threshold value; and
a distance image obtaining step of obtaining a distance image corresponding to said captured image by acquiring distance data of each pixel in said captured image, wherein:
the contour model deforming step includes deforming the contour model based on the obtained distance image, so that the contour of said target object is detected.
13. A computer readable storage medium storing a contour detecting program for making a computer execute an operation of detecting a contour of at least one target object in a captured image, the operation comprising:
a contour model deforming step of contracting or expanding a contour model consisting of a plurality of nodes, which surrounds at least one target object included in a captured image and is used for detecting a contour of said target object, by shifting the nodes based on a predetermined rule for deformation;
a split-determining step of calculating a distance between two non-adjacent nodes among the nodes of the contour model which was deformed in the contour model deforming step, and determining that the contour model is to be split when the calculated distance is equal to or smaller than a predetermined threshold value; and
a contour model splitting step of splitting the contour model according to a result of the determination executed in the split-determining step, wherein the splitting operation is executed in the vicinity of said two non-adjacent nodes.
14. The computer readable storage medium as claimed in claim 13 , wherein the operation further comprises:
a distance image obtaining step of obtaining a distance image corresponding to said captured image by acquiring distance data of each pixel in said captured image, and wherein:
the contour model deforming step includes deforming the contour model based on the obtained distance image, so as to make the contour model correspond only to said target object which is present in a predetermined distance range; and
the split-determining step includes determining whether the contour model defined in the predetermined distance range is to be split.
15. The computer readable storage medium as claimed in claim 13 , wherein the split-determining step includes calculating the square of the distance between said two non-adjacent nodes and determining that the contour model is to be split when the calculated square of the distance is equal to or smaller than the predetermined threshold value.
16. The computer readable storage medium as claimed in claim 13 , wherein the contour model deforming step includes deforming the contour model in a manner such that an energy function assigned to the contour model has a minimum solution.
17. The computer readable storage medium as claimed in claim 13 , wherein when a thinnest portion of said target object is known, the threshold value is predetermined in consideration of the thickness of the thinnest portion, so as to prevent the contour model from being split at the thinnest portion.
18. A computer readable storage medium storing a contour detecting program for making a computer execute an operation of detecting a contour of at least one target object in a captured image, the operation comprising:
a contour model deforming step of contracting or expanding a contour model consisting of a plurality of nodes, which surrounds at least one target object included in a captured image and is used for detecting a contour of said target object, by shifting the nodes based on a predetermined rule for deformation;
a calculating step for calculating a distance between two non-adjacent nodes among the nodes of the contour model;
a determining step for determining that the contour model is to be split when the calculated distance is equal to or smaller than a predetermined threshold value; and
a distance image obtaining step of obtaining a distance image corresponding to said captured image by acquiring distance data of each pixel in said captured image, wherein:
the contour model deforming step includes deforming the contour model based on the obtained distance image, so that the contour of said target object is detected.