Continuous monitoring of plant sub-cellular structural changes for plant and crop diseases detection by use of Intelligent Laser Speckle Classification (AI) technique

TL;DR

This paper presents an innovative AI-based laser speckle classification method for early, continuous, and cost-effective detection of plant diseases by analyzing sub-cellular structural changes.

Contribution

It introduces a novel combination of laser physics, texture analysis, and Bayesian classification for sub-cellular disease detection in plants.

Findings

High accuracy in distinguishing healthy and diseased plants.

Capability for continuous online monitoring via wireless network.

Potential to replace costly traditional field diagnostics.

Abstract

The continuous online monitoring of early signs of plant and crop diseases, at their early stages before a potential spread, is of high importance and necessitates multi-disciplinary techniques. Within this study a proposed technique achieves this goal by exploiting laser physics, textural image analysis, and AI for Shot hole disease. In this technique, specific laser light with a wavelength shorter than a sub-cellular component of an inspected plant, produces an interaction within the sub-cellular components and generates laser speckle patterns which can characterize those specific plant cells' features. The generated laser speckle image data then be quantized by texture analysis and classified by Bayesian networks. Such comparative methods manage to detect the differences at sub-cellular scales, such as nuclei modification, cellular shape, or size deformation, etc. for Shot hole disease with high classification accuracy between the healthy and diseased plants. The technique is capable of continuous online observation and monitoring of the plant or crop diseases via a wireless network at low instrumental cost and may replace the costly ground-truth field works