This pull request introduces a comprehensive pest detection feature, aimed at enhancing agricultural analysis capabilities. The feature is built around a custom-trained YOLOv8n model, specifically designed for high accuracy in pest detection and classification by @adityasngpta

Key Components Introduced:

• PestDetection class: A core component that loads the YOLOv8n model and performs pest detection on given images, returning annotated images with bounding boxes and class labels.
• /pest-detection endpoint: A new API endpoint that accepts image URLs for pest detection, processes them through the PestDetection class, and returns annotated images.
• Integration with Ingest class: The pest detection feature is seamlessly integrated into the existing Ingest class, allowing for easy invocation of pest detection alongside other functionalities like uploading annotated images to AWS S3

Technical Details:

• The YOLOv8n model is loaded from a specified path, with weights obtained from a dedicated Dropbox link.
• The pest detection process involves running inference on the provided images and extracting annotated images with detailed pest information.
• Annotated images are temporarily stored and uploaded to an S3 bucket, with URLs returned to the requester.

Usage:
This feature can be accessed via the /pest-detection endpoint, expecting a JSON payload with an image_urls key. It supports processing multiple images in a single request, enhancing efficiency and usability for end-users.

Future Directions:
While this implementation marks a significant advancement in our backend's capabilities, future work will focus on optimizing model performance, expanding the range of detectable pests, and improving the user experience with more detailed annotations.

This feature is expected to significantly contribute to the agricultural sector by providing advanced tools for pest detection and management, ultimately leading to better crop health and yield.