Anomaly detection using non-target clustering

1. A system for anomaly detection using non-target clustering, comprising:

a hierarchy generator comprising at least a plurality of programming instructions stored in a memory and operating on a processor of a computing device and configured to receive an image and to algorithmically process the image to produce a hierarchical image model of image information within the image, the hierarchical model comprising a plurality of pixel groupings produced and organized using an alpha-tree modeling algorithm;

a non-target clustering engine comprising at least a plurality of programming instructions stored in a memory and operating on a processor of a computing device and configured to process an image hierarchy to identify a background region and a plurality of anomalies, the anomalies being based at least in part on defined search criteria; and

a constraint connected component processing engine comprising at least a plurality of programming instructions stored in a memory and operating on a processor of a computing device and configured to process the contents of detected anomalies to identify the targeted objects based on defined search criteria;

wherein the alpha-tree modeling algorithm comprises a process for modeling pixel groupings within an image as hierarchically-nested image regions, the image regions being organized according to a dissimilarity metric.

2. The system of claim 1 , wherein the background region is identified by analyzing large-coverage components of the alpha-tree model.

3. The system of claim 2 , wherein the plurality of anomalies comprises at least a hole in the background region.

4. The system of claim 1 , wherein the image comprises an ocean region.

5. The system of claim 4 , wherein the background region comprises the ocean surface.

6. The system of claim 5 , wherein edge erosion is applied to filter out waves in the ocean surface.

7. The system of claim 1 , wherein the image comprises a land region.

8. The system of claim 7 , wherein the background region comprises the land surface.

9. The system of claim 8 , wherein spectral analysis is applied to filter out vegetation on the land surface.

10. A method for anomaly detection using non-target clustering, comprising the steps of:

receiving, at a hierarchy generator, an image;

processing the image using an alpha-tree modeling algorithm to produce a hierarchical model of image information within the image, the hierarchical model comprising a plurality of pixel groupings;

processing, using a non-target clustering engine, the hierarchical image model to identify a background region and a plurality of anomalies, the anomalies being based at least in part on defined search criteria; and

processing, using a constraint connected component processing engine, the contents of detected anomalies to identify the targeted objects based on defined search criteria;

wherein the alpha-tree modeling algorithm comprises a process for modeling pixel groupings within an image as hierarchically-nested image regions, the image regions being organized according to a dissimilarity metric.

11. The method of claim 10 , wherein the background region is identified by analyzing large-coverage components of the alpha-tree model.

12. The method of claim 11 , wherein the plurality of anomalies comprises at least a hole in the background region.

13. The method of claim 10 , wherein the image comprises an ocean region.

14. The method of claim 13 , wherein the background region comprises the ocean surface.

15. The method of claim 14 , wherein edge erosion is applied to filter out waves in the ocean surface.

16. The method of claim 10 , wherein the image comprises a land region.

17. The method of claim 16 , wherein the background region comprises the land surface.

18. The method of claim 17 , wherein spectral analysis is applied to filter out vegetation on the land surface.