Plant disease detection using image processing (MATLAB)

Published on 02 Jul 2021. Written by Palvi Soni

Nowadays plants are suffering many diseases due to widespread use of pesticides and sprays but identifying rotten areas of plants in the early stage can save plants. Examination of plants disease literally means examining various observable pattern on plants. Manually detecting disease in plants can be a tiresome process, hence image processing can do wonders in this context. Plant disease can be seen in different parts like in stem, root, shoot and even in fruit. Detection of plant disease by the automatic way not only reduces time but also it is able to save the plant from the disease in the beginning stage itself. We use different image processing techniques to predict the problem in plants.

Read more..

SLNOTE


Programming environment

MATLAB- MATLAB is used for the implementation of this project. There is a default image processing box in MATLAB that can enhance the working of the system. There are mathematical tools that provide image contrast and image intensity. In a nutshell, MATLAB is equipped with various features that support full image processing and image classification techniques.

We deployed MATLAB 2019 version for this purpose.

SLLATEST
Project implementation

We use image processing techniques in MATLAB. Firstly, we need plant image dataset and after that, all processing is done on that. We detect plant disease first and then we classify it using various image processing techniques. We use a neural network in this. The working of this project is as under:

• Acquisition of Plants images: We can obtain plant images from dataset of disease and we can also capture images manually by taking photographs of plants.
• Preprocessing of Plants images: It involves basic steps like image cropping, image smoothening, image snipping, image enhancement etc. This steps basically removes the extra noise of images.
• Extraction of various features: Feature extraction mainly emphasis on extracting relevant features and discarding irrelevant ones. It aims to enhance overall accuracy and help in speedy execution of training dataset.
• Training plants images: We perform training of the images for enhancing its tenacity or we can also say resolution.
• Segmentation of Plants images: Image segmentation means fragmenting images into parts so as to achieve better output. We will use threshold method in segmentation to achieve our result. Algorithm used in this is boundary and spot detection algorithm. After applying threshold algorithm we will also use Otsu method for segmentation.
• Neural Network classification: Artificial neural network has wonderful capabilities for performing classification, it plays a great role in recognizing patterns. It supports differentiation of classes depending upon similar groups in one class and non-similar in other ones. So simply in this project, it collects the diseased plants in one group and healthy plants in the other group.
• Disease detection: We can achieve this only after artificial neural network classification but this involves detecting outliers and separating them into the category, we mean to say detecting disease in plants.

SLDYK
Advantages of plant disease detection project:

1. It generates high accuracy results
2. It saves time
3. It produces efficient results.
4. It enhances plant productivity

Kit required to develop Plant disease detection using image processing (MATLAB):

• No kit required

Technologies you will learn by working on Plant disease detection using image processing (MATLAB):

• MATLAB
• Image processing
• CNN

Any Questions?