Pulse-coupled neural network performance for real-time identification of vegetation during forced landing

Abstract

Safety concerns in the operation of autonomous aerial systems require safe-landing protocols be followed during situations where the mission should be aborted due to mechanical or other failure. This article presents a pulse-coupled neural network (PCNN) to assist in the vegetation classification in a vision-based landing site detection system for an unmanned aircraft. We propose a heterogeneous computing architecture and an OpenCL implementation of a PCNN feature generator. Its performance is compared across OpenCL kernels designed for CPU, GPU, and FPGA platforms. This comparison examines the compute times required for network convergence under a variety of images to determine the plausibility for real-time feature detection. References

• Altera. Implementing FPGA design with the OpenCL standard. White Paper, Altera Inc., November 2012. http://www.altera.com/literature/wp/wp-01173-opencl.pdf
• Bittware. S5-PCIe-HQ user guide. Technical report, Bittware, Inc., September 2013.
• M. T. DeGarmo. Issues concerning integration of unmanned aerial vehicles in civil airspace. Technical Report mP 04W0000323, MITRE Corporation, 2004. https://www.mitre.org/sites/default/files/pdf/04_1232.pdf
• R. Eckhorn, H. J. Reiboeck, M. Arndt, and P. W. Dicke. A neural network for feature linking via synchronous activity: Results from a cat visual cortex and from simulations. In Models of Brain Function, pages 255–272, Cambridge, UK, 1989. Cambridge University Press.
• Khronos OpenCL Working Group. The OpenCL specification. Technical report, Khronos Group, October 2009. http://www.khronos.org/registry/cl/specs/opencl-1.0.pdf
• X. Gu, Y. Fang, and Y. Wang. Attention selection using global topological properties based on pulse coupled neural network. Computer Vision and Image Understanding, 117:1400–1411, 2013. doi:10.1016/j.cviu.2013.05.004
• X. Gu, L. Zhang, and D. Yu. General design approach to unit-linking pcnn for image processing. In Proceedings of International Joint Conference on Neural Networks, pages 1837–1841, 2005. doi:10.1109/IJCNN.2005.1556159
• Z. Li, R. F. Hayward, R. A. Walker, and Y. Liu. A biologically inspired object spectral-texture descriptor and its application to vegetation classification in power-line corridors. IEEE Geoscience and Remote Sensing Letters, 8(4):631–635, 2011. doi:10.1109/LGRS.2010.2098391
• A. Lu, W. Ding, J. Wang, and H. Li. Automonmous vision-based safe area selection algorithm for UAV emergency forced landing. In Proceedings of the International Conference on Information and Computer Applications 2012, pages 254–261, 2012. doi:10.1007/978-3-642-34041-3_37
• L. Mejias and P. Eng. Controlled emergency landing of an unpowered unmanned aerial system. Journal of Intelligent and Robotic Systems, 70:421–435, 2013. doi:10.1007/s10846-012-9767-5
• L. Mejias and D. L. Fitzgerald. A multi-layered approach for site detection in uas emergency landing scenarios using geometry-based image segmentation. In Proceedings of the 2013 International Conference on Unmanned Aerial Systems, pages 366–372, Atlanta, Georgia, 2013. IEEE Control Society. doi:10.1109/ICUAS.2013.6564710
• L. Mejias, D. L. Fitzgerald, P. C. Eng, and L. Xi. Forced landing technologies for unmanned aerial vehicles : towards safer operations. In Aerial Vehicles, pages 415–442, Kirchengasse, Austria, 2009. In-Tech. doi:10.5772/6481
• US-OSoD. Unmanned systems integrated roadmap fy2011-2036. Technical report, Office of the Secretary of Defense, US, 2011. http://info.publicintelligence.net/DoD-UAS-2011-2036.pdf
• Z. Wang, Y. Ma, F. Cheng, and L. Yang. Review of pulse-coupled neural networks. Image and Vision Computing, 28:5–13, 2010. doi:10.1016/j.imavis.2009.06.007