AI-Powered Plant Disease Detection System

Publication ID: 24-11856881_0001_PTD

Published: October 26, 2025

Category:Direct Improvements & Enhancements

Legal Citation

pr1or.art Inc., “AI-Powered Plant Disease Detection System,” Published Technical Disclosure No. 24-11856881_0001_PTD, Published October 26, 2025, available at https://archive.pr1or.art/24-11856881_0001_PTD

This technical disclosure describes improvements that would be readily apparent to a Person Having Ordinary Skill In The Art (PHOSITA) when considered in combination with the foundational architecture disclosed in U.S. Patent No. 11,856,881.

Background and Problem Solved

The original patent, 'Detection of plant diseases with multi-stage, multi-scale deep learning', demonstrated the effectiveness of deep learning in plant disease detection. However, it had limitations in terms of accuracy, scalability, and adaptability. The present invention addresses these limitations by introducing a hierarchical classification approach, online learning capabilities, and robust feature extraction methods, enabling more efficient and accurate plant disease detection.

Novelty and Inventive Step

The invention's novelty lies in the combination of hierarchical classification, online learning, and robust feature extraction, which enables more accurate, efficient, and adaptable plant disease detection. The inventive step is the integration of these components to overcome the limitations of the original patent.

Alternative Embodiments and Variations

Alternative embodiments of the invention include using different deep learning architectures, such as convolutional neural networks (CNNs) or recurrent neural networks (RNNs), or incorporating additional data sources, such as weather data or soil moisture levels, to further improve the accuracy of disease detection. Variations of the invention may also include using the system for detecting diseases in other types of plants or crops.

Potential Commercial Applications and Market

The invention has significant commercial potential in the agricultural industry, particularly in precision agriculture and crop monitoring. The system can be integrated into existing farm management systems, enabling farmers to detect diseases earlier and take targeted actions to prevent yield loss. The market potential is substantial, with the global precision agriculture market projected to reach $10.2 billion by 2025.

CPC Classifications

Section	Class	Group
A	A01	A01B79/005
G	G06	G06F18/214
G	G06	G06F18/24317
G	G06	G06F18/254
G	G06	G06N3/045
G	G06	G06N3/08
G	G06	G06T3/40
G	G06	G06T7/0012
G	G06	G06V10/764
G	G06	G06V10/774
G	G06	G06V10/82
G	G06	G06V20/188
G	G06	G06V20/60
G	G06	G06V20/68
A	A01	A01G7/00
G	G06	G06T2207/20016
G	G06	G06T2207/20081
G	G06	G06T2207/20084

Section 103 Obviousness Analysis (PHOSITA)

▼

Field of Art

Machine learning and computer vision for agricultural diagnostics, specifically deep learning techniques for plant disease detection and classification, involving multi-scale image processing and hierarchical classification models

Person of Ordinary Skill (PHOSITA) Profile

A skilled practitioner with expertise in deep learning, computer vision, agricultural image analysis, neural network architectures, and machine learning model development, typically holding a graduate degree in computer science, electrical engineering, or agricultural informatics

Obviousness Rationale

A person having ordinary skill would recognize that the PTD's disclosed variations represent predictable extensions of the source patent's multi-stage deep learning approach for plant disease detection. The proposed modifications leverage known machine learning techniques like online learning, multi-scale processing, and data augmentation to incrementally improve the original patent's classification methodology. These variations represent standard engineering approaches to enhancing machine learning models in image classification domains.

Obvious Combinations & Variations

Source Patent Element

Multi-stage classification model for detecting plant diseases using digital image processing

PTD Variation

Incorporating online learning capabilities to adapt to new disease types and variations

Obviousness Reasoning

Adapting machine learning models to handle evolving datasets is a well-established technique in the field, representing a predictable optimization of the original classification approach

Source Patent Element

First and second digital models for classifying image regions

PTD Variation

Using generative adversarial networks to augment training datasets with synthetic images

Obviousness Reasoning

Data augmentation through synthetic image generation is a known technique for improving machine learning model robustness and generalization

Source Patent Element

Image resizing and region extraction techniques for disease classification

PTD Variation

Multi-scale image processing to capture disease features at different resolution levels

Obviousness Reasoning

Implementing multi-scale analysis represents a standard approach to improving feature extraction and classification accuracy in computer vision applications

Source Patent Element

Deep learning model for classifying plant disease images

PTD Variation

Incorporating additional data sources like weather and soil moisture to enhance diagnostic accuracy

Obviousness Reasoning

Integrating contextual environmental data is a predictable extension of machine learning diagnostic systems to improve predictive performance

Source Patent Element

Computer system for applying classification models to plant disease images

PTD Variation

Expanding disease detection approach to different plant types and crop varieties

Obviousness Reasoning

Generalizing machine learning models across related domains represents a standard technique for expanding technological applicability

35 U.S.C. § 103 Summary: Based on the teachings of US Patent 11856881 and the disclosed variations, a person having ordinary skill in the art would find the proposed plant disease detection techniques obvious and non-inventive. The incremental modifications represent predictable applications of known machine learning methodologies to the existing multi-stage classification approach, thereby rendering potential patent claims obvious under 35 U.S.C. § 103.

Original Patent Information

Patent Number	US 11,856,881
Title	Detection of plant diseases with multi-stage, multi-scale deep learning
Assignee(s)	CLIMATE LLC