Description

Introduction

Deep learning is one of the subcategories of Artificial Intelligence. By using Deep learning object recognition can be deployed to classify and locate the object such as Person, Table, chair, plant, etc. It will be merged with robotics to follow the person.

---

Abstract

In this project, USB Camera is interfaced with the Raspberry Pi. Deep learning is applied to identify the person from the frame. And with the help of Computer Vision, Human follower application is developed based on the position of the person.

---

Existing system

In the existing system, Human follower application is done using sensors. Based on the distance between the robot and the person. The vehicle gets moved.

---

Proposed System

In this proposed system, Deep learning with Computer vision is used by obtaining the vision-based application using the camera.

---

Demo Video

---

Connection Description

In this project, the web camera is interfaced with the Raspberry pi through USB. The car will be operated by connecting the motors with the Driver through GPIO pins of the Raspberry Pi.

---

Project Description

This project initially opens the camera, by applying deep learning person will be classified and computer vision estimates the position of the person in the camera frame. Then based on the position of the person, Vehicle will be driven using Raspberry Pi to follow the human.

---

Hardware Required

• Raspberry Pi
• 16 GB SD card
• Camera
• Power supply
• LI – PO battery
• Driver circuit
• Robot setup

---

Software Required

• Raspbian OS with libraries installed
• SD Card Formatter
• Etcher/Win32 Disk imager

---

Result

In this project, Raspberry Pi with the camera to perform Deep learning is used to identify the person. Which has less FPS. To deploy the Human following application with a high frame rate with high efficiency, a Neural compute stick can be interfaced with the Raspberry Pi for high performance.