DETECTION AND CLASSIFICATION OF FRUIT DISEASES USING IMAGE PROCESSING
- Version
- Download 10
- File Size 574.18 KB
- File Count 1
- Create Date 1 May 2023
- Last Updated 1 May 2023
DETECTION AND CLASSIFICATION OF FRUIT DISEASES USING IMAGE PROCESSING
A.Bharathi1, E.Ashok kumar2 ,M.Srinithi3, B.Hariprakash4, M.Prabakaran5,
1Associate Processor in Information Technology, Nandha Engineering College, Perundurai -638052, Erode District, TamilNadu, India.
2,3,4,5Students of Information Technology, Nandha Engineering College, Perundurai -638052, Erode District, TamilNadu, India.
Abstract. Fruit diseases are important in agriculture worldwide. In this project, a neural network based on image processing is proposed to detect passion fruit disease. According to the CNN algorithm, the fruit image details are extracted from the first row used in this project by the existing package. However, it may take some time. Thus, the proposed system can be used for rapid and automatic identification of fruit diseases.
The proposed approach consists of the following main steps, including input image acquisition, image processing, affected area detection, affected area labeling, training set validation, and output display. Several types of fruit diseases such as bitter rot, puffball and powdery mildew are used for this approach. This approach was tested based on the type of fruit disease and the new and affected stages.
This algorithm is used to determine the type of fruit disease. Images are provided as sharp decay images, soft color images, and dust decay images. Before processing the image, it is converted to a color model to find the most suitable color model for this approach. Local binary patterns are used to extract features and support erosion methods are used to generate models. According to this approach, fruit diseases can be identified with an average accuracy of 79% and stages with an average accuracy of 66%.
Download
