Abstract
To meet the real-time requirements of major industry players and society, today, far and wide, there are video surveillance systems. With the recent development in technology and reduction in hardware costs, the number of cameras installed per kilometer is now increasing. The major challenge with video analytics is storage and response time. This chapter discusses transforming an IP camera into an Intelligent Camera by proposing a hybrid method of computational intelligence techniques like fuzzy, genetic, swarm optimization, reinforcement learning, ensemble methods, and deep belief networks to perform analysis at the place of data generation itself. With the acquired intelligence, a hybrid algorithm for anomaly detection and scene identification using fuzzy logic and deep learning can be designed. Deep learning models promise quick response time and better accuracy levels on image identification.
TopTraditional Camera Vs Smart Camera
There is a great difference between the traditional camera and the smart camera. Both have the same purpose collecting data for analysing but smart camera proves to be better at the field. The smart camera consumes less labour, time and memory. Figure 1 shows the beta version of converting a traditional IP camera into an intelligent and smart computing device.
Figure 1. Traditional IPC to Smart Camera
Key Terms in this Chapter
Smart Camera: A camera that has additional built-in computational hardware and deciding algorithms.
Machine Learning: A branch of Artificial Intelligence that learns from data and automates pattern recognition, image identification and decision making with minimum human intervention.
5G: Fifth generation high speed and high bandwidth wireless communication technology.
Anomaly: Anomaly is a person or thing or situation that deviates from normal appearance or happening.
Edge Computing: A networking and computing method to perform analytics and computation at the place of data generation to reduce latency.
Neural Network: A set of algorithms to identify the relationship between collection of data available like human brain cognition.