Development of Security System for ATM with the Aid of Computer Vision

Nowadays ATMs are seen in more number of places like, near hospitals, near shopping malls, near educational institutions, near bus stops, etc. Automated Teller Machines (ATMs) are part of most of our lives which eases access to financial transactions without direct interaction with bank authorities. As the number of ATM installations are increasing, threats related to burglary in ATMs are also increasing. In current work we have developed a security system for ATMs which locks the door based on the abnormal actions performed by the consumer. Our proposed system deals with embedded system to lock or unlock the door depending upon the abnormal behaviour of the person inside ATM cabin by implementing Haar cascade classifier to detect human body and the video encountered in abnormal human behaviour will be sent to nearby police station and nearby bank by establishing Client-Server communication using Socket programming.