PLANT DISEASE IDENTIFICATION using AIML and Data science

PLANT DISEASE IDENTIFICATION
Is using a AIML and Database to store the images.

1. PLANT DISEASE IDENTIFICATION FOR EFFECTIVE CROP MANAGEMENT GUIDE : MRS. SOWNTHARYA.C (AP/CSE) TEAM MEMBERS MUTHUKKUMARAN.B BALASHANKAR.M ANNAMALAI.S

2. AGENDA Introduction Background of Problem Problem definition Literature Review Implementation plans Implementation status Publication Status

3. Introduction Deep learning is a subfield of artificial intelligence focused on training algorithms called neural networks to learn from data and make predictions or decisions. These neural networks are structured like the human brain, with interconnected nodes (neurons) that process and analyze information. Through deep learning, these networks can automatically learn features and patterns from large datasets, enabling tasks like image and speech recognition, natural language processing, and even complex tasks such as autonomous driving or medical diagnosis. The key to its power lies in its ability to discover intricate patterns and representations within data, leading to highly accurate and adaptable models.

4. Background of the problem Complexity of Plant Diseases: Plant diseases can be caused by various factors such as pathogens (bacteria, fungi, viruses), pests, environmental conditions (temperature, humidity), nutrient deficiencies, and more. Identifying the specific disease affecting a plant requires expertise in plant pathology and a deep understanding of the symptoms, causal agents, and influencing factors. Impact on Crop Yield and Quality: Untreated plant diseases can lead to significant crop losses, reduced quality of agricultural products, and economic hardship for farmers. Timely and accurate diagnosis is crucial to implement appropriate management strategies and prevent further spread of diseases.

5. Problem definition Agriculture plays a crucial role in global food security and economy. However, one of the major challenges faced by farmers worldwide is the occurrence of plant diseases. Plant diseases can lead to significant yield loss, economic burden, and food shortages. Timely detection and management of these diseases are essential to ensure crop health and productivity.

6. Literature Review 1) “Real-Time Detection Of Crop Leaf Diseases Using Enhanced Yolov8 Algorithm” Houda Orchi , Mohamed Sadik ,  Mohammed Khaldoun ,  Essaid Sabir (2023) The identification of factory conditions in real- time presents a significant challenge for growers, particularly in regions like Morocco where husbandry is vital to the frugality. Limited coffers and moxie frequently hamper accurate opinion, leading to wasted time and sweats in trying to regain affected crops. In response to this pressing issue, a new approach using recent advancements in computer vision and deep literacy technology has been proposed. This innovative system utilizes the YOLOv8 algorithm, famed for its effectiveness in object identification, to dissect splint images at a rapid-fire rate of 70 FPS( Frames Per Second).

7. Literature Review By employing a grid- grounded analysis fashion and a single neural network, this system enhances the delicacy and effectiveness of complaint identification, offering a timely result for growers. The effectiveness of YOLOv8 in this environment surpasses traditional object identification algorithms, reaffirming its status as a commanding result for real- world operations. This groundbreaking exploration not only demonstrates the eventuality of advanced technology in husbandry but also underscores the significance of invention in addressing critical challenges faced by growers worldwide.

8. Literature Review 2)"Detection of Plant Diseases in an Industrial Greenhouse: Development, Validation & Exploitation" Yassine Lakhdari , Enric Soldevila, Jihene Rezgui , Éric Renault (2023) In our study, we concentrate on the discovery of conditions in tomato shops within similar settings. Our primary objects revolve around the development and confirmation of a discovery system exercising the YOLOv8 model, with an emphasis on its practical connection. To support our exploration, we curated a unique dataset containing images of tomato leaves tormented by colorful conditions, serving as a precious resource for training and assessing our discovery model. Employing the YOLOv8 armature, a slice- edge object discovery frame, we conducted expansive trials to assess its performance in directly relating diseased areas on tomato leaves.

9. Literature Review Our trial encompassed the disquisition of different parameters, including different training strategies, data addition ways, and hyperparameter configurations, furnishing perceptivity into optimal settings for achieving high discovery delicacy and robustness. Also, we validated the practical mileage of our developed model by enforcing it within a real- life artificial hothouse terrain. This demonstration showcased the flawless integration of our discovery system into functional workflows, pressing its eventuality to prop hothouse drivers in early complaint discovery, monitoring, and decision- making processes. Our first results demonstrate promising complaint discovery on tomato leaves under hothouse conditions. Our stylish model scores0.8 for mAP50. While there are areas of enhancement, these early results demonstrate the great eventuality of our methodology to ameliorate complaint operation in agrarian surroundings.

10. Implementation plans This Flask project allows users to upload images and receive predictions about the contents of those images using a YOLO (You Only Look Once) model. HTML Templates: index.html: This template provides a simple form where users can upload images. result.html: This template displays the predictions made by the YOLO model for the uploaded image . Flask App (app.py):Flask: Initializes the Flask application.

11. Implementation plans SQLAlchemy : Handles the database operations. The Image model represents the uploaded images. YOLOv8: Initializes the YOLOv8 model for object detection. METHODOLOGIES INSIDE THE PREDICTION : confidence_thresholds : Dictionary defining confidence thresholds for each class label. index(): Renders the index.html template, providing the upload form. upload_image (): Handles the image upload process. Upon receiving a POST request with an image file, it reads the image data, saves it to the database, and redirects to the prediction route.

12. Implementation plans predict(): Retrieves the latest uploaded image from the database, performs object detection using the YOLO model, filters the predictions based on confidence thresholds, and renders the result.html template with the predictions. Additionally, it saves the annotated image with bounding boxes drawn around detected objects to a file. Explanation: When a user uploads an image, it is saved to the SQLite database ( new.db ) as binary data using SQLAlchemy . The predict() function retrieves the latest uploaded image from the database, decodes it, and passes it to the YOLO model for object detection.Object detection results are filtered based on confidence thresholds defined in confidence_thresholds . Predictions exceeding the threshold are displayed to the user.Annotated images with bounding boxes drawn around detected objects are saved to the file system and displayed to the user along with the predictions.

13. Implementation plans Error Handling: If there are no predictions available, the user is informed accordingly. Error messages, such as 'list' object has no attribute 'boxes', indicate issues with accessing attributes or methods on objects and require debugging to resolve.Overall , the project demonstrates the integration of Flask, SQLAlchemy , and a YOLO model for image uploading, object detection, and prediction display.

14. Implementation status Firstly, the index.html will appear to choose the files from device.

15. Implementation status Then we select the .jpeg file as input file.

16. Implementation status After upload it gets predict and redirect to the result.html file to show the predicted result.

17. Implementation status This image shows the images are stored in the database of SQLAlchemy . It stores after uploading and it retrives to the prediction page.

18. Publication Status

19. THANK YOU……