امنیت با استفاده از پردازش تصویر و شبکه عصبی پیچشی عمیق Security using Image Processing and Deep Convolutional Neural Networks

I. INTRODUCTION Technology used in securing highly important places has changed a lot since the last few years and will continue to change in the coming years. Security is very important when it comes down to smart applications. The new and emerging concept of smart security offers a convenient, comfortable, and safe way for securing highly sensitive areas [1]. Security systems used conventionally aim to protect a place from a breach by sending a notification in the form of a triggered alarm at the time of breach. However, the proposed security system offers many more benefits when compared to the conventional systems which are discussed in detail as we go further ahead into the implementation and working of this system. This paper focuses on how security at locations considered very sensitive and private such as a location where highly valuable or sensitive data is stored can be made much more effective by deploying intelligent systems that are capable of performing with efficiency levels that cannot be achieved by a human or even other traditional security systems. This system comprises of two modules defined at the hardware level which includes a Raspberry Pi Microcontroller with a few sensors connected to it and an Arduino Microcontroller with Global System for Mobile Communications (GSM) and Global Positioning System (GPS) capabilities installed together at the area of deployment. These two modules communicate with each other on the local network and together communicate with the users on the remote public network. This paper also focuses on how computer vision can be used to detect human activity at the site and the use of deep convolutional neural networks to identify and match an image with a set of people authorized to visit a site. Thus, with a combination of both, the efficiency of the system is increased multifold.