Around 70% of Indian economy is reliant on horticulture. Malady on plants, models substantial decrement in both the quantity and quality of horticulture products. Hence, detection and diagnosis of disease at the apparent time is crucial to the cultivator. Farmers find it cumbersome to analyze whether a plant is diseased or not manually as it requires a lot of time, labour and cost. Thus, it is always required to automate a system to identify if a plant is diseased or not earlier to evolve. This paper actualizes a machine learning system to recognize the type of malady on various plant species where phases include dataset acquisition, feature extraction, training and classification. The datasets of both healthy and diseased leaves are trained using various machine learning classifiers. Feature extraction is done using Histogram of Oriented Gradient (HOG) where features of color, shape and texture are extracted. Then, the accuracy of all classifiers is compared to obtain the best classifier. Random Forests is found to be the best classifier for all variety of plants. Later, a test image of leaf is supplied to the classifier to classify whether the leaf is diseased or not and retrieve the disease type.
A Novel Method for Plant Leaf Malady Recognition using Machine Learning Classifiers
2019-06-01
1553538 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Engineering Index Backfile | 1931
|Predicting Red Light Running Violation Using Machine Learning Classifiers
Springer Verlag | 2021
|