Team Data - Decoders Mid Pro

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

Plant Disease

Detection
By Team Data_decoders
Importance of
Plant Disease
detection
Plant disease detection aligns with sustainable agriculture
principles by reducing chemical inputs.
Accurate disease detection enables informed decisions on
disease control measures.
Effective disease detection protects agricultural systems,
prevents crop losses, and contributes to global food security.
Problem
Statement
Develop a plant disease recognition
system for accurate identification of
diseases in various plant species.
Analyse plant images to determine if
they are healthy or infected with a
specific disease.

LEARN MORE

Work FLow

Results
Exploring Data Model
Preparation Modelling evaluation and
the data
insights

Tools To Be Used

Programming Language: Python 02


01 Deep Learning Framework: PyTorch

01 Data Manipulation and Analysis: NumPy,


Data Visualization: Matplotlib,
03 Pandas 04 Seaborn

Model Training and Evaluation:


05 Image Preprocessing: OpenCV
06 PyTorch, Scikit-learn
03

Development Environment: Jupyter Cloud Deployment: AWS (Amazon Web


Notebook, Anaconda Services), or other cloud platforms for
07 08

future deployment
Importing
Libraries

os: For working with files


numpy (np): For numerical computations
pandas (pd): For working with dataframes
torch: PyTorch module
matplotlib.pyplot (plt): For plotting
information on graphs and images using
tensors
torch.utils.data.DataLoader: For creating
dataloaders
Importing
Libraries

PIL.Image: For checking images


torchvision.utils.make_grid: For visualizing
data
torchvision.transforms: For transforming
images into tensors
torchvision.datasets.ImageFolder: For
working with classes and images
%matplotlib inline: Enables inline plotting in
Jupyter notebooks
Exploring the
Data
Dataset: "New Plant Diseases Dataset
(Augmented)"
The dataset consists of 87,000 Images
Split: 80% of the images in the training set,
20% in the validation set
Disease classes: Total of 38 disease types
Plants: Dataset includes 14 different plant
types
Diseases: 26 disease types (excluding
"healthy" class)
Image quality: Varies across the dataset,
with variations in clarity and lighting levels
observed
Exploring the data
Data
Preparation
ImageFolder: Automatically maps folder
structure to class labels
transforms.ToTensor(): Converts images to
tensors and scales pixel values to the range
of 0 to 1
DataLoader: Efficiently loads and
preprocesses data in batches during model
training
GPU Acceleration: Utilizes GPU for faster
computations and improved performance
Helper Functions: Facilitate the movement
of data from CPU to GPU for running the
notebook on GPU
Data Preparation
Sample of Batch Size
Data
Preparation
Since our dataset is very large and it is very
computationally hard and time taking
process to run it on cpu, so we use GPU to
accelerate our calcuations, and give
computational power to our model. But to
run out notebook on GPU we need to move
our data(variables, objects) from CPU to
GPU, and to do so we will write helper
functions.

Visit the link


Thank You

You might also like