# BOLLWM: A real-world dataset for bollworm pest monitoring from cotton fields in India

**Authors:** Jerome White, Chandan Agrawal, Anmol Ojha, Apoorv Agnihotri, Makkunda Sharma, Jigar Doshi  
**arXiv:** [2304.00763](https://arxiv.org/abs/2304.00763)  
**Date:** 2023-04-03  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2304.00763v1/

## Key points

- The paper presents a unique dataset of agricultural pest images captured over five years by thousands of smallholder farmers and farming extension workers across India.
- The dataset has been used to support a mobile application that relies on artificial intelligence to assist farmers with pest management decisions.
- The creation of the dataset was achieved through a combination of organized data collection and less controlled mobile application usage, making it distinct within the pest detection community.
- The characteristics of the dataset place it closer to other non-agricultural object detection datasets, which opens up opportunities for a wide variety of research agendas beyond just pest management applications.
- This work provides valuable insights into how such datasets can be created and utilized in real-world settings, particularly in developing countries where smallholder farmers are prevalent.
- It highlights the potential impact technology can have on improving crop yields and livelihoods for farmers.
- Overall, this work contributes to advancing the field of computer vision for agriculture by providing an example of how AI-driven solutions can be implemented in practical settings.

## Summary

The paper "BOLLWM: A real-world dataset for bollworm pest monitoring from cotton fields in India" presents a unique dataset of agricultural pest images captured over five years by thousands of smallholder farmers and farming extension workers across India. The dataset has been used to support a mobile application that relies on artificial intelligence to assist farmers with pest management decisions. The creation of the dataset was achieved through a combination of organized data collection and less controlled mobile application usage, making it distinct within the pest detection community. The characteristics of the dataset place it closer to other non-agricultural object detection datasets, which opens up opportunities for a wide variety of research agendas beyond just pest management applications. This paper provides valuable insights into how such datasets can be created and utilized in real-world settings, particularly in developing countries where smallholder farmers are prevalent. It also highlights the potential impact technology can have on improving crop yields and livelihoods for farmers. Overall, this work contributes to advancing the field of computer vision for agriculture by providing an example of how AI-driven solutions can be implemented in practical settings.