Scholarlink

Context <p>Wheat is one of the most common staple crops in world and contributes to food security worldwide. As per the National Agricultural Institute, about 3.6% of wheat grain quantity has been reduced due to various fungal, bacterial, and virus-based diseases. Thus, early prediction of diseases is essential through timely control measures that minimize grain yield losses.</p> Objective <p>The main objective of this paper is to present the comprehensive review and analyze studies that have been published between 1997 and 2024, which shows the image processing procedures for wheat disease (WD) recognition.</p> Methods <p>Five research questions are proposed and this review analyze 169 studies that has been finalized through systematic literature review (SLR) approach. Further, this study summarized image preprocessing methods for infected leaf area segmentation along with their localization and classification techniques.</p> Results <p>A total number of 169 studies have been published in 43 reputed journals, 44 conferences that have been identified. After analysis five different types of image segmentation techniques were analyzed. After analysis, the Mask-RCNN instance segmentation technique achieves a 98.81% localization rate for FHB disease in wheat spikes. Additionally, the combination of Logistic Regression and hue moment features with ResNet-50 achieves 99.8% classification accuracy for yellow rust WD.</p> Conclusion <p>Throughout this review, ROI generation instance segmentation methods aim to find and draw the bounding boxes of infected wheat plants that are still in infancy. This survey aims to inspire readers to delve into image segmentation-based models and encourage comprehensive exploration of their applications in this field.</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

Impact of image segmentation and feature sets in automated plant disease classification: a comprehensive review based on wheat plant images

  • 期刊论文

  • 发表日期：

    2025年8月13日

查看全文
  • Deepak Kumar,
  • Vinay Kukreja

摘要

Context

Wheat is one of the most common staple crops in world and contributes to food security worldwide. As per the National Agricultural Institute, about 3.6% of wheat grain quantity has been reduced due to various fungal, bacterial, and virus-based diseases. Thus, early prediction of diseases is essential through timely control measures that minimize grain yield losses.

Objective

The main objective of this paper is to present the comprehensive review and analyze studies that have been published between 1997 and 2024, which shows the image processing procedures for wheat disease (WD) recognition.

Methods

Five research questions are proposed and this review analyze 169 studies that has been finalized through systematic literature review (SLR) approach. Further, this study summarized image preprocessing methods for infected leaf area segmentation along with their localization and classification techniques.

Results

A total number of 169 studies have been published in 43 reputed journals, 44 conferences that have been identified. After analysis five different types of image segmentation techniques were analyzed. After analysis, the Mask-RCNN instance segmentation technique achieves a 98.81% localization rate for FHB disease in wheat spikes. Additionally, the combination of Logistic Regression and hue moment features with ResNet-50 achieves 99.8% classification accuracy for yellow rust WD.

Conclusion

Throughout this review, ROI generation instance segmentation methods aim to find and draw the bounding boxes of infected wheat plants that are still in infancy. This survey aims to inspire readers to delve into image segmentation-based models and encourage comprehensive exploration of their applications in this field.