Enhancing Sugarcane Crop Health: CNN and SVM-Based Predictive Analysis of Leaf Diseases

This research provides an extensive assessment of the classification performance for seven different types of leaf diseases in sugarcane. Important metrics for each illness class are listed in the accompanying table, including precision, recall, F1-score, support, and total accuracy. With precision...

Full description

Saved in:
Bibliographic Details
Published in2024 3rd International Conference for Innovation in Technology (INOCON) pp. 1 - 6
Main Authors Singh, Shivanshi, Banerjee, Deepak, Chauhan, Rahul, Joshi, Kireet, Singh Gill, Kanwarpartap
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.03.2024
Subjects
Adaptation models
agricultural
crop
Crops
diseases
precision
Predictive models
Productivity
Reliability
Sustainable development
Technological innovation
Online AccessGet full text

Cover

More Information
Summary:This research provides an extensive assessment of the classification performance for seven different types of leaf diseases in sugarcane. Important metrics for each illness class are listed in the accompanying table, including precision, recall, F1-score, support, and total accuracy. With precision rates ranging from 94.34% to 94.86%, this thorough evaluation shows a consistent model performance and a high degree of accuracy in correctly recognizing positive instances within each illness category. Recall values also agree with these precision rates, confirming that the model is good at identifying genuine positive cases and reducing false negatives. The model's resilience in diagnosing sugarcane leaf diseases is demonstrated by the F1-score, which constantly maintains a value of 94.57% while effectively harmonizing precision and recall. Support metrics, which range from 795 to 875, show the frequency of recurrence of each class in the sample, giving crucial information. These values highlight the model's versatility across a wide range of disease occurrences by providing insights into the relative prevalence of each type of disease. The model's overall accuracy, which stands at 94.57%, demonstrates its ability to reliably forecast all classes. It is the percentage of correctly classified cases in the dataset. This study makes a substantial contribution to the field of agricultural disease identification and has applications that could improve crop management and increase agricultural output. Our classification model's strong performance makes it a potentially useful tool for early disease detection in agricultural settings, which might help sugarcane growers all over the world avoid suffering large losses.
DOI:10.1109/INOCON60754.2024.10511876