Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

An Organic Weed Control Prototype using Directed Energy and Deep Learning

Deng Cao, Hongbo Zhang, Rajveer Dhillon

TL;DR

This paper addresses organic weed control by merging a directed-energy robotic platform with deep learning-based weed recognition. It develops two prototypes (Phase I and Phase II) and builds soybean and sweet corn image databases to train transfer-learned CNNs that achieve up to about 98% accuracy in classifying weed versus crop under natural field conditions. The integrated system aims to perform real-time, selective, non-chemical weed eradication, with plans to shorten treatment time and enhance autonomy. The work advances sustainable farming by demonstrating high-accuracy crop/weed discrimination and proposing a practical, non-toxic energy-based weed-control workflow for small organic farms.

Abstract

Organic weed control is a vital to improve crop yield with a sustainable approach. In this work, a directed energy weed control robot prototype specifically designed for organic farms is proposed. The robot uses a novel distributed array robot (DAR) unit for weed treatment. Soybean and corn databases are built to train deep learning neural nets to perform weed recognition. The initial deep learning neural nets show a high performance in classifying crops. The robot uses a patented directed energy plant eradication recipe that is completely organic and UV-C free, with no chemical damage or physical disturbance to the soil. The deep learning can classify 8 common weed species in a soybean field under natural environment with up to 98% accuracy.

An Organic Weed Control Prototype using Directed Energy and Deep Learning

TL;DR

This paper addresses organic weed control by merging a directed-energy robotic platform with deep learning-based weed recognition. It develops two prototypes (Phase I and Phase II) and builds soybean and sweet corn image databases to train transfer-learned CNNs that achieve up to about 98% accuracy in classifying weed versus crop under natural field conditions. The integrated system aims to perform real-time, selective, non-chemical weed eradication, with plans to shorten treatment time and enhance autonomy. The work advances sustainable farming by demonstrating high-accuracy crop/weed discrimination and proposing a practical, non-toxic energy-based weed-control workflow for small organic farms.

Abstract

Organic weed control is a vital to improve crop yield with a sustainable approach. In this work, a directed energy weed control robot prototype specifically designed for organic farms is proposed. The robot uses a novel distributed array robot (DAR) unit for weed treatment. Soybean and corn databases are built to train deep learning neural nets to perform weed recognition. The initial deep learning neural nets show a high performance in classifying crops. The robot uses a patented directed energy plant eradication recipe that is completely organic and UV-C free, with no chemical damage or physical disturbance to the soil. The deep learning can classify 8 common weed species in a soybean field under natural environment with up to 98% accuracy.
Paper Structure (12 sections, 3 equations, 8 figures)

This paper contains 12 sections, 3 equations, 8 figures.

Table of Contents

  1. Introduction
  2. Related Work
  3. Approach/Methodology
  4. Prototype Phase 1
  5. The Phase II prototype
  6. Development of Soybean and Corn Databases
  7. The Soybean Database
  8. The Sweet Corn Database
  9. Network architecture
  10. Results
  11. Conclusion
  12. Acknowledgments

Figures (8)

  • Figure 1: Left: The front view of the DA. Right: The bottom view of the DA.
  • Figure 2: The Demonstration of DA at WPAFB test site.
  • Figure 3: The DA towed by a WC800-DM4, 4WD All Terrain Robot.
  • Figure 4: The proposed prototype Disruptor Array Robot version 4 (DAR4)
  • Figure 5: The end effector of the robotic arm of DAR4.
  • ...and 3 more figures