Scholarlink

Background and objective <p>Wheat rust diseases are caused by the Puccinia fungus pathogen and are one of the most common threats in wheat production. The rust diseases encompass stem, stripe, and leaf rust variants, each posing a significant threat to both yield quality and grain quantity. The rust diseases are identified by farmers who have no experience, which leads to incorrect identification, resulting in significant losses.</p> Objective <p>The rust diseases are identified by farmers with no experience leads to incorrect identification. A computer vision-based automatic system (CVBA) is important for rust disease identification.</p> Method <p>This paper proposes a computer vision-based novel model, namely as Adaptive Simple K-means Clustering and Fuzzy-based Pyramid Support Feature model (ASK-FPS), consisting of two phases. Firstly, the local gradient ratio pattern of each pixel in the wheat leaf image is extracted as a feature. Then superpixel is generated using Adaptive superpixel linear iterative clustering (ASLIC) and k-means clustering. In second phase, the pyramid of the histogram of oriented gradient (PHOG) features extracts the superpixel features that are reduced by Principal component analysis (PCA), and Analysis of variance (ANOVA), which is further applied to the fuzzy support vector machine (FSVM) model.</p> Results and conclusion <p>The proposed model selects cumulative variance preservation (95%), reducing the feature space to 10 principal components and achieves a maximum accuracy of 98.89%, precision of 96.49%, recall of 95.59%, F1-score of 95.74%, specificity of 98.42% and Matthew’s coefficient of 0.952 with FSVM classifier. The automated ASK-FPS model helps to identify healthy and rust leaves, including healthy, stem, stripe, and leaf rust diseases, accurately.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Adaptive computer vision based k means clustering and fuzzy pyramid support feature model for accurate classification of wheat rust disease

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月13日

查看全文
  • Deepak Kumar,
  • Vinay Kukreja,
  • Vandana Ahuja,
  • Manmohan Sharma,
  • Vivek Bhardwaj

摘要

Background and objective

Wheat rust diseases are caused by the Puccinia fungus pathogen and are one of the most common threats in wheat production. The rust diseases encompass stem, stripe, and leaf rust variants, each posing a significant threat to both yield quality and grain quantity. The rust diseases are identified by farmers who have no experience, which leads to incorrect identification, resulting in significant losses.

Objective

The rust diseases are identified by farmers with no experience leads to incorrect identification. A computer vision-based automatic system (CVBA) is important for rust disease identification.

Method

This paper proposes a computer vision-based novel model, namely as Adaptive Simple K-means Clustering and Fuzzy-based Pyramid Support Feature model (ASK-FPS), consisting of two phases. Firstly, the local gradient ratio pattern of each pixel in the wheat leaf image is extracted as a feature. Then superpixel is generated using Adaptive superpixel linear iterative clustering (ASLIC) and k-means clustering. In second phase, the pyramid of the histogram of oriented gradient (PHOG) features extracts the superpixel features that are reduced by Principal component analysis (PCA), and Analysis of variance (ANOVA), which is further applied to the fuzzy support vector machine (FSVM) model.

Results and conclusion

The proposed model selects cumulative variance preservation (95%), reducing the feature space to 10 principal components and achieves a maximum accuracy of 98.89%, precision of 96.49%, recall of 95.59%, F1-score of 95.74%, specificity of 98.42% and Matthew’s coefficient of 0.952 with FSVM classifier. The automated ASK-FPS model helps to identify healthy and rust leaves, including healthy, stem, stripe, and leaf rust diseases, accurately.