초록 ▼

A method of target recognition performs a 3D comparison of target and reference data. Translation invariant signatures are derived from the two data sets, and an estimate of the orientation of the target with respect to the reference is obtained. Rotational alignment and comparison can then be achie

A method of target recognition performs a 3D comparison of target and reference data. Translation invariant signatures are derived from the two data sets, and an estimate of the orientation of the target with respect to the reference is obtained. Rotational alignment and comparison can then be achieved. The 3D data sets can be represented on an axi-symmetric surface such as a sphere and rotational convolution, over a discrete set of selected rotation angles can be performed. Optic flow can be used to derive the estimate of orientation or the target relative to the reference, in terms of a displacement field.

대표청구항 ▼

1. A method of target recognition comprising the steps of: providing a set of 3D data corresponding to a reference object;obtaining a set of 3D data corresponding to a target object;deriving a translation invariant 3D signature for each of the 3D data sets;comparing the translation invariant 3D sign

1. A method of target recognition comprising the steps of: providing a set of 3D data corresponding to a reference object;obtaining a set of 3D data corresponding to a target object;deriving a translation invariant 3D signature for each of the 3D data sets;comparing the translation invariant 3D signatures for the reference object and the target object to derive an estimate of the orientation of the target with respect to the reference data;bringing the target data and reference data into rotational alignment according to said derived estimate of orientation; andcomparing the aligned target data with the reference data and deriving a measure of similarity based on said comparison. 2. A method according to claim 1, wherein the 3D target data and the 3D reference data are represented on an axisymmetric surface. 3. A method according to claim 2, wherein the 3D target data and the 3D reference data are represented on the surface of a sphere. 4. A method according to claim 1, wherein the translation invariant 3D signature of a data set is derived from a modulus of the 3D transform of that data set. 5. A method according to claim 4, wherein the transform is a Fast Fourier Transform. 6. A method according to claim 1, wherein comparison of the 3D target data and the 3D reference data comprises 3D rotational convolution. 7. A method according to claim 1, wherein comparison of the 3D target data and the 3D reference data is performed over a discrete set of selected rotation angles. 8. A method according to claim 7, wherein a first subset of angles are selected varying only about a first axis. 9. A method according to claim 8, wherein said first axis is selected based on an estimate of a local terrain gradient. 10. A method according to claim 1, wherein said 3D data corresponding to said target comprises a plurality of defined 3D points. 11. A method according to claim 10, wherein each point is measured as range and direction values relative to a given viewpoint. 12. A method according to claim 1, wherein said 3D data corresponding to said target is obtained by 3D laser imaging. 13. A method according to claim 1, further comprising comparing the rotationally aligned 3D target data and the 3D reference data to derive an estimate of a position of target with respect to the reference data; bringing the target data and the reference data into translational alignment according to said estimate of position. 14. A method according to claim 1, wherein the derived estimate of the orientation of the target with respect to the reference data is refined by calculating an optic flow between the 3D target data and the 3D reference data. 15. A method according to claim 14, wherein calculating the optic flow comprises rotationally aligning the target data and the reference data according to an initial estimate of orientation;deriving a translation invariant signature for the rotationally aligned data sets; anddetermining the displacement field between the signatures. 16. A method according to claim 1, wherein hidden surface removal is applied to a reference model. 17. A method according to claim 1, wherein a correlation score is obtained for the estimated orientation, and hidden surface removal is employed selectively in dependence upon said correlation score.