PhamIQ

AI-Powered Crop Disease Detection for Ghana's Farmers. Multispectral anlysis and personalised AI chat system."From Farm Fields to Healthy Yields - AI That Sees What You Can't See"

Comment 2

Our Story: From Soil to Software

We didn’t start out writing code. We started with calloused hands and dirt under our fingernails.

I’m Alex. My father still wakes up at 5 AM to tend to his yam, cassava, and maize fields in the Oti Region. Silas’s mother grows the best tomatoes in Ashanti, and Asante's family has been growing cashews in Brong-Ahafo for generations. Farming isn’t just part of our story — it’s in our blood.

We’re the kind of team that traded hoes for keyboards but never forgot where we came from.

"You cannot solve what you do not understand. We understand farming because farming is in our DNA." – Our team motto

The 5 AM Heartbreak

Imagine your father checking his maize field before sunrise. One day, he spots yellow patches on the leaves. Is it a disease? A nutrient deficiency? A pest?

This is the reality for most farmers in Ghana:

  • Walk hours to the nearest extension office.
  • Wait days or weeks for an expert to respond.
  • Lose up to 50% of their crops while waiting.
  • Spend life savings on the wrong treatment.
  • Cannot determine whether their lands are suitable for the crops they grow.
  • Cannot monitor certain environmental conditions of their farms, such as temperature, vegetation health, soil type, and many others.

We've watched this happen to our families. Entire harvests lost — not because farmers didn't care, but because help came too late.

When the Ghana AI Hackathon challenged us to build a smart agriculture solution, we knew this wasn’t just another coding project.

What Phamiq Does

Phamiq is our answer — a platform powered by AI and multispectral imaging that helps farmers diagnose crop diseases quickly and accurately.

Diagnose Crop Diseases

  • Upload a photo of your crop
  • Get an instant diagnosis with a confidence score
  • Works with both standard images and multispectral images for early, precise detection

Multispectral Image Analysis

After detecting crop diseases, it's crucial for farmers to understand the environmental factors influencing them.

Our app uses satellite imagery and multispectral data to give farmers deep insights into their land covering:

  • Temperature
  • Vegetation health
  • Soil type
  • Atmospheric moisture
  • And more

We analyse data from platforms like Landsat, **Sentinel-2, and **DJI drones, capturing light beyond the visible spectrum (RGB, NIR, SWIR). This enables us to uncover vital land and crop health information the human eye can't see.

Key indices we calculate:

  • NDVI – Vegetation health
  • LST – Surface temperature
  • SAVI/EVI – Detailed vegetation analysis
  • NDBI/BSI – Built-up & bare soil detection
  • NDMI – Vegetation moisture

Using Analytical Hierarchy Process (AHP), we intelligently combine these factors to generate a **Crop Suitability Index (0–1). A score closer to **1 means the land is ideal for crops like cashew, **cassava, **maize, or **tomatoes.

We don’t just help you detect disease — we help you understand your land, optimize your crops, and make smarter decisions.

Supports Multiple Crops

  • Cassava, cashew, maize, tomato
  • Automatically selects the right model for the crop

Built-in AI Assistant

  • Ask questions about diseases, treatments, or platform usage
  • Get responses tailored to local farming needs

Learn & Discover

  • Explore disease information, symptoms, and recommended actions

Track History

  • Every diagnosis and conversation is saved
  • Review past results and insights anytime

Secure & Easy to Use

  • Account-based access
  • Data is private, secure, and only visible to the user

How We Built It

Phamiq is a full-stack AI platform, built with farmers in mind but powered by serious engineering.

AI Models

  • Trained deep learning models (CNNs) using **PyTorch
  • Converted to ONNX for fast, real-time inference
  • Developed custom preprocessing pipelines for multispectral images

Backend

  • Built with FastAPI (Python)
  • Modular architecture: image processing, multispectral pipeline, AI inference, chat, user management
  • Uses MongoDB to store diagnoses, chats, and user data
  • Dockerized for easy deployment and scalability

Frontend

  • Built with React, **TypeScript, and **Tailwind CSS
  • Clean, responsive, and mobile-friendly
  • Handles image uploads, results, chat, and history in real time

Challenges We Faced

  1. Making Multispectral Work
  • Processing complex image channels and training robust models took time and experimentation
  1. Real-Time Inference
  • ONNX optimization and backend architecture had to balance speed and accuracy, especially with large images
  1. Smooth Integration
  • Ensured seamless communication between the frontend, backend, and AI model
  1. Non-Tech-Friendly UX
  • Designed everything to be usable even by someone who’s never touched a computer before
  1. Data Collection
  • Getting high-quality, labeled images — especially multispectral ones — wasn’t easy, but we did it

What We’re Proud Of

  • Built an end-to-end disease detection tool: from image upload to diagnosis
  • Advanced multispectral image support: detects early-stage diseases others miss
  • Accessible interface: built for farmers, not just researchers
  • Secure and scalable architecture: ready for real-world deployment
  • Team synergy: we built Phamiq through passion, shared purpose, and lived experience

What We Learned

  • How to build real-time, multispectral AI tools for non-technical users
  • That tech solutions must be empathetic, not just functional
  • The power of combining deep learning, human experience, and thoughtful design

What’s Next for Phamiq

  • Digital Pest & Disease Alert Network Establish a real-time alert system across Ghana to help farmers respond faster to outbreaks.

  • Government & Agri-Cooperative Integration Collaborate with agri-cooperatives and support national programs like the Planting for Food & Jobs Initiative.

  • Offline-Capable Mobile App Launch a lightweight mobile app that works in remote and low-data areas — 60% of farms currently have limited connectivity.

  • Crop & Regional Expansion Extend Phamiq’s support to include more crop types and coverage across all 16 regions of Ghana.

  • Local Language Support Integrate local languages to ensure accessibility for all farmers, regardless of literacy or language barriers.

  • Community-Driven Platform Foster connections between farmers, extension officers, and agricultural experts to encourage shared learning, collaboration, and trust.

Final Word

Phamiq is more than an app. It’s a promise to the people who raised us. It’s our way of saying:

"We see you. We remember. And now, we’re building something better."

Built With

Share this project:

Updates