Full PDF of Dissertation


Abstract: 

This project aims to combine various ideas about computer vision, specifically region detection, object recognition, and change detection, with real world applications. High Spy is intended to be a system that is comprised of Unmanned Aerial Vehicles, otherwise known as UAVs, for the purposes of collecting data from the surrounding landscape. These UAVs will collect video data and return it to a base where the rest of the system will use various methods to analyse the data to create a knowledge base about the surrounding landscape. Once this is done, the system will continue to analyse video returned by the UAVs to look for changes in the landscape from mission to mission. It is the hope that this will lead to a military application at a Forward Operating Base, or FOB. In this way, the UAVs will look for Improvised Explosive Devices, or IEDs, alerting a squadron commander if anything is found.

 


Footage Provided by:

DraganFly Innovations Inc

MarcusUAV

References: 

[1] J.L. Johnson and D. Ritter. Observation of periodic waves in a pulse-coupled neural network. Opt. Lett., 18(15):1253–1255, Aug 1993.
[2] R. Eckhorn, H.J. Reitboeck, M. Arndt, and P. Dicke. Feature linking via synchronization among distributed assemblies: Simulations of results from cat visual cortex. Neural Computation, 2(3):293–307, Fall 1990.
[3] J.L. Johnson and M.L. Padgett. Pcnn models and applications. Neural Networks, IEEE Transactions on, 10(3):480 –498, may 1999.
[4] M. Isard and A. Blake. Condensation—conditional density propagation for visual tracking. International Journal of Computer Vision, 29:5–28, 1998.
[5] C.J. Munno, H. Turk, J.L. Wayman, J.M. Libert, and T.J. Tsao. Automatic video image moving target detection for wide area surveillance. In Security Technology, 1993. Security
Technology, Proceedings. Institute of Electrical and Electronics Engineers 1993 International Carnahan Conference on, pages 47 –57, oct 1993.
[6] V. Saligrama, J. Konrad, and P. Jodoin. Video anomaly identification. Signal Processing Magazine, IEEE, 27(5):18 –33, sept. 2010.
[7] D. M. Greig, B. T. Porteous, and A. H. Seheult. Exact maximum a posteriori estimation for binary images. Journal of the Royal Statistical Society. Series B (Methodological),
51(2):pp. 271–279, 1989.
[8] L. R. Ford and D. R. Fulkerson. Maximal flow through a network. Canadian Journal of Mathematics, 8:399–404, 1956.
[9] Y. Boykov, O. Veksler, and R. Zabih. Fast approximate energy minimization via graph cuts. Pattern Analysis and Machine Intelligence, IEEE Transactions on, 23(11):1222 –1239, nov 2001.
[10] V. Kolmogorov and R. Zabin. What energy functions can be minimized via graph cuts? Pattern Analysis and Machine Intelligence, IEEE Transactions on, 26(2):147 –159, feb. 2004.
[11] Y. Boykov and V. Kolmogorov. An experimental comparison of min-cut/max- flow algorithms for energy minimization in vision. Pattern Analysis and Machine Intelligence, IEEE Transactions on, 26(9):1124 –1137, sept. 2004.
[12] Y. Boykov and G. Funka-Lea. Graph cuts and efficient n-d image segmentation. International Journal of Computer Vision, 70:109–131, 2006. 10.1007/s11263-006-7934-5.
[13] G. Funka-lea, Y. Boykov, C. Florin, M. p. Jolly, R. Moreau-gobard, R. Ramaraj, and D. Rinck. Automatic heart isolation for ct coronary visualization using graph-cuts. 2006.
[14] S. Lloyd. Least squares quantization in pcm. Information Theory, IEEE Transactions on, 28(2):129 – 137, mar 1982.
[15] J. MacQueen. Some methods for classification and analysis of multivariate observations. In Proc. Fifth Berkeley Sympos. Math. Statist. and Probability (Berkeley, Calif., 1965/66), pages Vol. I: Statistics, pp. 281–297. Univ. California Press, Berkeley, Calif., 1967. 
[16] S. Bagon. Matlab wrapper for graph cut. http://www.wisdom.weizmann.ac.il/bagon, December 2006.