Skip to main content

Home / Projects / CropOptima - Crop Disease Detection App

CropOptima - Crop Disease Detection App

Top 50 national Bangkit Academy Android app for crop disease detection using Kotlin, Machine Learning, backend APIs, and VPS deployment.

Role

Android Developer & Cloud Engineer

Duration

6 months

Team

7 developers

CropOptima - Crop Disease Detection App — landscape view
CropOptima - Crop Disease Detection App — portrait view

Tech Stack

Kotlin · Machine Learning · Backend API · VPS · Android

CropOptima is an Android application built during Bangkit Academy 2023 Batch 2 to detect crop disease using Machine Learning, with backend APIs and model deployment handled on a VPS for real-time inference.

Problem

Petani membutuhkan cara yang lebih cepat untuk mengenali indikasi penyakit tanaman. Pemeriksaan manual bisa lambat dan tidak selalu mudah dilakukan saat keputusan di lapangan harus cepat.

Solution

Membangun aplikasi Android berbasis Kotlin yang memakai Machine Learning untuk membantu deteksi penyakit tanaman, dengan backend API dan model deployment di VPS agar inference bisa berjalan secara real-time.

Architecture Decisions

  • Chose Kotlin for native Android development and direct access to Android platform capabilities
  • Deployed ML model and backend APIs on a VPS for real-time inference
  • Kept the mobile app and inference backend separated through API boundaries
  • Focused the product flow on disease detection rather than broad farming features

Trade-offs

  • Native Android over cross-platform — faster delivery for the Bangkit team scope, but iOS was not supported
  • Server-side inference over on-device inference — easier model deployment, but depends on network access
  • Focused disease detection scope over broad farm management features

Lessons Learned

  • ML model accuracy heavily depends on training data quality
  • Real agricultural data is noisy — need robust error handling
  • Farmers prefer simple UI over feature-rich interfaces
  • Agile methodology critical for 7-person team coordination

Features

  • Machine Learning-based crop disease detection
  • Android app built with Kotlin
  • Backend API integration
  • VPS-based model and API deployment
  • User authentication (register, login, reset)
  • Detection result visualization
  • Responsive and intuitive UI

Challenges Overcome

  • Deploying ML models and backend APIs on a VPS
  • Handling real-time inference flow between Android and backend services
  • Designing a user-friendly mobile interface for agricultural users
  • Coordinating delivery in a multi-person Bangkit Academy capstone team

Achievement

  • Top 50 National Teams, Bangkit Academy 2023 Batch 2