COTTONACE
Problem
Challenges in Pest Management for Cotton Farmers
Pest-related crop loss is a major challenge for millions of cotton farmers across India. Infestations like the pink bollworm can destroy up to 70% of cotton yields if not detected and managed early.
High crop losses due to late pest detection
Excessive pesticide use harms environment and health
Rising input costs and declining productivity
Lack of real-time, localized advisory for farmers
Why Solving This Matters
Cotton is India’s largest cash crop and a livelihood source for over 6 million farmers. Without timely intervention, pest attacks like pink bollworm can devastate yields and incomes, especially in low-resource rural areas. CottonAce enables early, expert-led decisions through AI-powered advisories, directly supporting national goals in agriculture, climate resilience, and rural development.
Protects farmer livelihoods
Reduces crop losses from pests like pink bollworm through early detection
Enables informed, real-time decisions
Localized, timely pest advisories empower farmers and extension officers
Optimizes pesticide usage
Cuts unnecessary spraying, lowering input costs and reducing environmental harm
Improves extension services
Gives field staff and administrators tools for better outreach and decision-making
Our Solution
A digital solution supporting farmers, field officers, and agri programs with AI-powered pest management.
Who it is For
For Farmers
Snap, analyze, and get instant pest management advice in your local language
For Extension Officers
Monitor fields, communicate effectively, and build stronger farmer relationships
For Program Teams
Real-time dashboards and data-driven insights for smarter interventions
What's in it for Farmers
Created by fajar hasyim
from the Noun Project
Snap & Send
Use pheromone traps and upload pest images via the CottonAce app
Instant Insights
AI detects pest type and infestation severity in real-time
Localized Advisories
Get timely, stage-specific spray recommendations in 9 Indian languages
Smart Features
Multilingual (9 languages)
Mandi prices, weather, and tutorials
Works offline
Direct chat with extension officers
For Extension Officers
Field-Level Monitoring
Verify pest alerts and follow up on advisories with farmers
Simplified Communication
Geo-specific support builds stronger, trusted relationships
Training & Knowledge Access
App-based modules help officers stay updated and train farmers better
For Program Teams & Admins
Real-Time Dashboard
Visualize pest outbreaks, spray trends, and farmer engagement across regions
Data-Driven Targeting
Insights inform smarter interventions and program design
Scalable & Modular
Easily adaptable across states, languages, or agriculture schemes
The Technology Behind CottonAce
AI Pest Detection
Image recognition for pink & American bollworms
Economic Threshold Logic
ICAR-CICR backed thresholds guide spray timing
Privacy & Inclusivity by Design
Consent-driven, role-based access for secure and ethical deployment
Multilingual Machine Learning
Inclusive, region-specific language training
Who can use it
Primary End Users
Cotton Farmers
Use the CottonAce mobile app to monitor pest levels, receive local advisories, and make timely decisions to protect their crops.
Extension Officers / Field Facilitators
Support farmers in trap setup, data collection, and interpreting advisories. Help verify infestations and provide on-ground guidance.
Other System Users
Agriculture Program Administrators
Use the real-time dashboard to monitor pest trends, program reach, and impact across regions. Improve planning and responsiveness
State & National Government Departments
Integrate CottonAce into public agriculture schemes to scale support and reduce crop loss in high-priority cotton-growing regions
NGOs and Farmer Producer Organizations (FPOs)
Deploy the solution to strengthen community-led pest management and improve engagement with smallholder farmers
AgriTech Platforms
Integrate CottonAce’s plug-and-play components (e.g., advisory engine, pest detection API) into broader farm management solutions
from the Noun Project
Researchers & Academic Institutions
Access anonymized, real-world pest and intervention data to study trends, improve ETL thresholds, and build future AI models
Policy Makers
Use aggregate insights to shape better pest management strategies, optimize subsidy allocation, and reduce environmental harm
Benefits
Improved Efficiency
AI automates pest detection and advisories, reducing manual effort for farmers and extension workers
Early Detection, Timely Action
Real-time, localized pest alerts help farmers act before damage escalates
Cost Savings
Enables targeted pesticide use, cutting unnecessary input costs and minimizing crop loss
Localized & Inclusive
Multilingual and offline-ready, ensuring access across regions and literacy levels
Better Decision-Making
Dashboards offer actionable insights for program teams and policymakers across regions and time
Scalable & Plug-and-Play
Modular and deployable across programs, geographies, and platforms as a digital public good
Performance Indicators
70% Crop Loss Prevented
Timely pest alerts helped reduce pink bollworm damage across monitored regions
9 Languages Supported
Inclusive advisory delivery tailored to diverse linguistic communities
1000s of Traps, Weekly Insights
Field data fuels hyperlocal, real-time decision-making for farmers
Smarter Pesticide Use
Less chemical usage, more trust—validated by on-ground field studies
Deployed at Scale
Integrated into state-level agriculture programs by government and NGO partners
Technical Architecture
Created by Prijun Koirala
from Noun Project
AI Models
Pest Detection Model
Deep learning model trained on pheromone trap images to identify pink and American bollworms.
ETL Logic Engine
Applies scientifically validated thresholds based on crop stage and farmer type (organic/chemical) to trigger advisories.
Translation & NLP Models
Supports multilingual UI through Indic NLP tools and machine translation.
Created by Vectplus
from the Noun Project
Data Pipelines
Field Data Collection
Images and metadata from the field collected via mobile app in 9 languages (offline-first).
Secure Processing & Sync
Data anonymized, synced, and processed on secure cloud services.
Dashboard Integration
Feeds into real-time dashboards to visualize pest trends and intervention outcomes.
Farmer App
Multilingual, offline-ready Android app for capturing pest data and receiving advisories tailored to crop stage and farmer type.
Extension Officer Portal
Admin interface for validating submissions, managing farmer records, and supporting timely field interventions
Monitoring Dashboard
Real-time web dashboard for program teams to track infestations, app usage, and intervention outcomes across geographies.
Technical Foundation
APIs & Integration Modules
Reusable APIs for pest classification, ETL engine, user management, and advisory generation.
Deployment Stack
Android OS for field deployment, cloud platforms for inference and analytics.
Open Source Tools & Dependencies
Annotation Tools – For labeling pest images used in model training AI4Bharat models for ASR and translation
Datasets – Image & advisory data, weather inputs, and localized pest info
Scientific Guidelines – Based on ICAR-CICR, CIB&RC protocols for ETL
Partner Organizations – NGOs, FPOs, government bodies aiding last-mile delivery
Weather APIs – Open weather APIs, advisories from ICAR-CICR website and resource materials, from PAU and other SAUs
How to Use
Pre-requisties
(Languages, libraries, system requirements)
For Mobile App
Android 8.0 or higher
Internet (periodic sync; works offline otherwise)
Camera-enabled device
For Admin Dashboard
Web browser (latest Chrome, Firefox, Edge)
Stable internet connection
Role-based login credentials
For AI & Integration Modules
Python 3.8+
Libraries: Pytorch,Pytorch Lightning, Tritonserver
GPU recommended for model retraining: Yes
Docker (for containerized deployments)
Usage Guide
Follow these steps to use the system
Using the Mobile App (Farmers & Extension Officers)
Download the CottonAce app from Play Store
Sign in with phone number (OTP-based)
Select your preferred language
Upload trap photos and receive advisory alerts
Admin Dashboard
Login with assigned credentials
View infestation heatmaps, farmer activity, and reports
Filter by geography, pest type, or time period
Contribution Guidelines
We welcome contributions! Please read our contribution guidelines before submitting PRs
How to Contribute:
Fork this repository
Create a feature branch (git checkout -b feature-xyz)
Make your changes and test thoroughly
Submit a pull request with clear documentation
Raise issues via the Issues tab for bugs, feature requests, or feedback
We encourage contributions to:
Improve pest detection accuracy
Add support for more languages
Optimize performance for low-resource devices
Integrate with new weather or agriculture data sources
Inner -Source Info
This project is licensed under the Apache License 2.0, a permissive inner-source license that allows commercial use, modification, distribution, and private use. It requires preserving copyright and license notices, grants contributors’ patent rights, and permits redistribution under different terms without mandating source code disclosure.
Contributors
Team or Contributors
Nevil Vekariya
Associate ML Scientist - II
Sudharshan Sekhar
Machine Learning Scientist
Contact Persons
Nevil Vekariya
Associate ML Scientist - II
Email ID:
community.kiran@wadhwaniai.org
Wadhwani AI @ 2025. All rights reserved.
