THE 3 rd INTERNATIONAL SCIENTIFIC CONFERENCES OF STUDENTS AND YOUNG RESEARCHERS dedicated to the 99
th
anniversary of the National Leader of Azerbaijan Heydar Aliyev
173
RASPBERRY PI BASED SYSTEM FOR VISUAL OBJECT DETECTION AND TRACKING Ramil Abbaszade Baku Higher Oil School Baku, Azerbaijan ramil.abbaszade.std@bhos.edu.az Supervisor: PhD, Associate Professor Leyla Muradkhanli Keywords: Python, Yolo algorithm, Raspberry Pi, OPENCV
Detecting and tracking objects is one of the most common and difficult
jobs that a monitoring system must do in evaluate relevant events and
suspicious actions, as well as automatically analyze and extract video
information. An object might be a face, a head, a human, a line of people, a
crowd, or a product on an assembly line, according to the business
intelligence concept.
Object detection is undergoing a fast revolution in the
field of computer vision. Its role in both object classification and object
localization makes it one of the most difficult problems in the field of computer
vision. The purpose of this detection approach is to establish where objects
are situated in a given image, a process known as object localization, and to
which category each item belongs, a process known as object classification.
In this paper, I will provide the reader with the most common trends in
terms of Visual Object Detection and Tracking. Object identification methods
include rapid R-CNN, Retina-Net, and Single-Shot MultiBox Detector (SSD).
Despite the fact that these techniques have overcome the constraints of data
limitation and modeling in object identification, they are not capable of
detecting objects in a single algorithm run. There are 8 most common
algorithms for object detection. The one which I will use is called YOLO (You
Only Look Once). This algorithm is well-known for its speed and precision. It
has been used to identify traffic signals, pedestrians, parking meters, and
animals in a variety of applications.
YOLO is implemented using OpenCV deep learning library in Python.
YOLO performs object detection as a regression problem and returns the
class probabilities of the collected images. To detect objects in real-time, the
YOLO adapted convolutional neural networks (CNN). To identify objects, the
approach requires only a single forward propagation through a neural
network, as the name implies. This indicates that the complete image is
predicted in a single algorithm run. The CNN is used to forecast several class
probabilities and bounding boxes at the same time. There are several
variations of the YOLO algorithm. Tiny YOLO and YOLOv3 are two popular
examples.
Advantages of YOLO