The proposed work introduces an algorithm to prevent collision between vehicles, inferring distance from digitally acquired real-time images. The digital image acquisition is aided by image sensors, mounted at the front of the vehicle. But, such a system proves fruitfulness only at a rapid rate of signal acquisition. A digital image being bulky further slows down this process. The challenge faced at high rate multi-dimensional signal acquisition from the Shannon sampling theorem hinders the real-time practice of this system. But, the paradoxical theory of compressive sensing breaks this challenge by contradicting the sampling theorem. This work makes use of block compressive sensing, taking minimum random measurements from a block divided pixel space. The reconstructed signal from these nonlinear measurements undergoes image processing to converge to a suitable action. The output is validated by using peak signal-to-noise ratio and structural similarity index.
P. Sreevidya, S. Veni, Rajeev, V., and Krishnanugrah, P. U., “Chapter 9 - Compressive Sensing-Aided Collision Avoidance System”, in The Cognitive Approach in Cloud Computing and Internet of Things Technologies for Surveillance Tracking Systems, D. Peter, Alavi, A. H., Javadi, B., and Fernandes, S. L., Eds. Academic Press, 2020, pp. 121-140.