AI-Based Plant Disease Detection and Recommendation System
Budget: ₹1,500 – ₹12,500 INR
I need a web-based solution that lets farmers upload a clear photo of a leaf or crop (JPEG or PNG only). As soon as the image lands on the server, an AI model should analyse it and immediately show, on the same page, the disease name, a short description, likely causes, treatment methods, prevention tips, suggested fertilisers or pesticides, and any extra crop-care recommendations. The whole idea is to help users decide what to do in the field without waiting for an agronomist.
A robust convolutional-network or comparable vision model is fine, as long as it reaches reliable accuracy on common regional crops. Training can be done with open datasets or a curated set I will supply once development starts. The interface must be clean, mobile-friendly, and fast enough to work over rural internet connections. No e-mail or SMS integration is required; all feedback stays on screen.
Deliverables
• Trained AI model with documented metrics
• Responsive web front-end for image upload and results display
• Back-end code (Python preferred, TensorFlow or PyTorch acceptable) with clear instructions to reproduce training and deploy the model
• Setup script or Dockerfile for one-click installation on an Ubuntu VPS
• A concise user guide and API reference
Source code ownership transfers fully to me at project hand-off, and I will test the system against an unseen image set before final acceptance.
A robust convolutional-network or comparable vision model is fine, as long as it reaches reliable accuracy on common regional crops. Training can be done with open datasets or a curated set I will supply once development starts. The interface must be clean, mobile-friendly, and fast enough to work over rural internet connections. No e-mail or SMS integration is required; all feedback stays on screen.
Deliverables
• Trained AI model with documented metrics
• Responsive web front-end for image upload and results display
• Back-end code (Python preferred, TensorFlow or PyTorch acceptable) with clear instructions to reproduce training and deploy the model
• Setup script or Dockerfile for one-click installation on an Ubuntu VPS
• A concise user guide and API reference
Source code ownership transfers fully to me at project hand-off, and I will test the system against an unseen image set before final acceptance.