Number Plate Detection and Extraction
• Primary Goal
: To develop a system that can detect and extract number plates from images.
•
250 Images for
objection detection
were collected and labelled using
Image Annotation Tool
-
LabelImg
.
•
Carried out data preprocessing, build and trained
deep learning object detection model
(
InceptionResnet V2
) in
Tensorflow
.
•
Extracted text from the images using
Optical Character Recognition API Tesseract
.
•
Built a
Pipeline Deep Learning model
and developed a web app using
Flask
.
•
Trained
YOLOv5
model with Custom data.