A Hybrid DenseNet201-SVM for Robust Weed and Potato Plant Classification
DOI:
https://doi.org/10.26555/jiteki.v8i2.23886Keywords:
DenseNet201, KNN, Potato, SVM, WeedAbstract
Potato plant growth needs to be protected from weeds that grow around it. Currently, the manual spraying of pesticides by farmers is not only precise on weeds but also on cultivated plants. Therefore, we need an intelligent system that can appropriately classify potato plants and weeds. The research contribution combines feature extraction and appropriate classification methods to obtain optimal accuracy. In addition, the small amount of data also contributes to this research. In this research, it is proposed to use a combination of feature extraction using deep learning techniques and classification using machine learning. We use the feature extraction method with the DenseNet201 model because this study's data is not too much. Complex vectors from DenseNet201 were reduced using Principal Component Analysis (PCA). Then we classified it with the Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) classification methods. The experimental results show that the PCA method can reduce the complexity of high-dimensional features into 2 and 3 dimensions. The average of the best classification results using SVM was obtained with a 3-dimensional PCA configuration, but on the contrary, using KNN obtained the best results in a 2-dimensional PCA configuration. The results showed 100% accuracy on the DenseNet201-SVM hybrid. The SVM kernel configuration used is a linear kernel. The results of this study can be an insight into an accurate classification method for separating weeds and potatoes so that agricultural technology can apply this method for classification.
Downloads
Published
Issue
Section
License
Authors who publish with JITEKI agree to the following terms:
- Authors retain copyright and grant the journal the right of first publication with the work simultaneously licensed under a Creative Commons Attribution License (CC BY-SA 4.0) that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work.
This work is licensed under a Creative Commons Attribution 4.0 International License