Issuu

International Research Journal of Engineering and Technology (IRJET)

e-ISSN: 2395-0056

Volume: 09 Issue: 05 | May 2022

p-ISSN: 2395-0072

www.irjet.net

Object Detection Bot Reshma R1, Sohan LV2, Vrushank RP3 ,Vivek Pavan U4 ,Dr Jyothi R5 1,2,3,4 Student,

Computer Science and Engineering, Global Academy of Technology, Karnataka, India 3Example: Associate Professor, Computer Science and Engineering, Global Academy of Technology, Karnataka, India. ---------------------------------------------------------------------***-------------------------------------------------------------------Abstract - Internet of Things (IoT) is an ever-expanding network of smart and functional items. For security and safety 5

applications, unambiguous object detection is critical so that no events are missed and no resources are wasted due to false alarms. The deep learning and IoT services are available on cloud platforms, which can be leveraged to build a fully automated and failsafe security solution. Safety and security have always been a requirement for urban civilization. By combining an application operating on an edge device with AWS cloud services, our research presents accurate object detection for indoor and outdoor settings. Our project's main goal is to save human labor by developing and deploying a safety and security system that automatically detects plastic bottles, the presence of metal cans, or abandoned baggage while utilizing IoT and cloud-based services to reduce setup costs. For the compute-light detection task, we discover that AWS Rekognition has a lower delay and is cost effective than all other cloud platforms. We provide results for identifying plastic bottles, detecting tin cans and monitoring any abandoned bags that might be a potential threat to society.

Key Words: Internet of Things, Object detection, Deep learning, AWS cloud services, AWS Rekognition, 1. INTRODUCTION Surveillance systems are placed at various public locations to prevent littering or possible threat from an abandoned bag. Owing to the advancement in technology, it is difficult to monitor the CCTV feeds all week, thus demanding the need for a robust security mechanism that alerts the necessary security personnel in-case of incidents of public littering or threats from abandoned baggage. The objective of this paper is securing outdoor environment by building an autonomous bot using a Raspberry Pi, sensors and camera modules. Cloud platforms have a worldwide presence, allowing data and applications to be accessible with minimal latency due to the physical proximity of cloud servers, which benefits both data producers and users. The Amazon Web Services (AWS) cloud platform is the most mature cloud platform with the most cloud services. AWS offers a variety of services that may be used in conjunction with IoT to create a complete smart security solution. The Amazon Web Services (AWS) cloud platform is the most well-known cloud platform with the most extensive variety of cloud services. AWS provides a number of services that may be used with IoT to build a complete smart security and safety solution. Images of bottles, baggage are stored in the cloud, and in the presence of plastic bottles or baggage is detected, the bot captures the image, send it to the AWS cloud, and compare that with hundreds of images in the database. When the image matches with the images in the database, the security personal is alerted and necessary actions are carried out.

2. LITERATURE REVIEW A few recent works on real-time object detection were recently reported. The report's contents are weighed in terms of their benefits and drawbacks. Cleanliness as well as safety at outdoors are traditionally led manually by humans or surveillance system, which are not that efficient as well as not accurate. There are a few smart systems on the market that use electronic sensors put on the cane, but these systems have several drawbacks. Using the notion of motion detection, the authors of [1] created an IoT-based smart surveillance system. The systems architecture consists of raspberry pi, a webcam, a GSM module and Wi-Fi module that detects, captures images and notifies user . This system suffers from certain disadvantages like inability to perform object detection , causing not only false positives but also not having the ability to exclude certain trusted detections . Nag et al. [2] proposed a facial recognition and detection algorithm-based security system for accessing doors. The authorized user receives the incoming data from the Raspberry Pi via the Telegram application. This approach relies on HaarCascades, an ancient method for detecting faces that takes a lot of memory. Hassan et al. [9] use face recognition to lock and open the door. To create security solutions, [4] and [5] employ Raspberry Pi and IoT sensors. The paper "Indoor Autonomous Vehicle Navigation—A Feasibility Study Based on Infrared Technology" [6 ] discusses how sensors may be used to motion and presence of humans, both of which are important for security. Karri and

Impact Factor value: 7.529

ISO 9001:2008 Certified Journal

Page 3309