Crop insect classification by combining RGB and Segmented images using SVM and KNN
Keywords:
Elbow Clustering, KNN, SVM, crop insects, image segmentation, image fusionAbstract
The Indian economy is badly affected if the crop production for a particular financial year goes down. The two major factors that affect crop production are environmental factors and the other is different types of crop disease caused by various insects. In the current work we have proposed a solution to correctly classify the various crop insect images with the help of fusing RGB and Segmented images. The elbow method has been adopted to correctly identify the number of clusters. The k-nearest neighbor has been used to extract the features by clustering the nearest pixel for segmenting the image, and combination of KNN with support vector machine has been used to fuse the RGB image and segmented image. This approach has streamed line our data and thus the fused image has been feed to pre-trained CNN model Resnet50. The classification accuracy has been observed as 85.5,72.7,88.4 in case of RGB, Segmented, and Fused image respectively. So the proposed method has achieved the highest accuracy, thus novelty has been achieved through the proposed methods.
Significant Statement: The work has been done for the classification of the crop insects, for the classification we have used the fusion of RGB and segmented images of insects. At first we have segmented the image using the K nearest neighbor algorithm. The no of cluster has been chosen with the help of elbow method. Then we have proposed an algorithm to fuse the RGB and segmented image. The novelty has been achieved as the RGB image provide the color information and the segmentation try to focus on the required object. Fusing both of them leads to better classification with the help of Resnet50.
Downloads
References
Tonnang Henri EZ, Sokame M Bonoukpoè, Abdel-Rahman M Elfatih, ubois T, 2022 Measuring and modelling crop yield losses due to invasive insect pests under climate change. Current Opinion in Insect Science 50:100873. https://doi.org/10.1016/-j.cois.2022.100873.
Showkat S, Qureshi S, 2022, Efficacy of Transfer Learning-based ResNet models in Chest X-ray image classification for detecting COVID-19 Pneumonia. Chemometrics and Intelligent Laboratory Systems 224:104534.
You L, Zheng L, Jiang J, Zhao Y, 2022, GPU-accelerated Faster Mean Shift with euclidean distance metrics. IEEE 46th Annual Computers, Software, and Applications Conference (COMPSAC): 211-216. https://doi.org/10.1109/COMPSAC54236.2022.-00037
Bosakova-Ardenska A, Kutryanska M, Boyanova P, Panayotov P, 2022, Application of images segmentation and median filter for white brined cheese structure evaluation. AIP Conference Proceedings 2570(1):020014. https://doi.org/10.1063/5.0099673.
Onumanyi AJ, Molokomme DN, Isaac SJ, Abu-Mahfouz AM, 2022, AutoElbow: An Automatic Elbow Detection Method for Estimating the Number of Clusters in a Dataset. Applied Science 12:7515. https://doi.org/10.3390/app12157515
Carneiro GA, Magalhães R, Neto A, Sousa JJ, Cunha A ,2022, Grapevine Segmentation in RGB Images using Deep Learning. Procedia Computer Science 196:101-106. https://doi.org/10.1016/j.procs.2021.11.078.
Chen C, Chandra S, Seo H, 2022, Automatic Pavement Defect Detection and Classification Using RGB-Thermal Images Based on Hierarchical Residual Attention Network. Sensors 22:5781. https://doi.org/10.3390/s22155781
Peignier S, Lacotte V, Duport MG, Baa-Puyoulet P, Simon JC, Calevro F, et.al ,2023, Detection of Aphids on Hyperspectral Images Using One-Class SVM and Laplacian of Gaussians. Remote Sensing 15:2103. https://doi.org/10.3390/rs15082103
Gomes L, Torres RdS, Côrtes ML, 2023, BERT- and TF-IDF-based feature extraction for long-lived bug prediction in FLOSS: A comparative study. Information and Software Technology 160:107217. https://doi.org/10.1016/j.infsof.2023.107217.
Chao Z, Duan X, Jia S, Guo X, Liu H, Jia F, 2022, Medical image fusion via discrete stationary wavelet transform and an enhanced radial basis function neural network. Applied Soft Computing 118:108542. https://doi.org/10.1016/j.asoc.2022.108542.
Ullah, H., Zhao, Y., Abdalla, F.Y.O., 2022, Fast local Laplacian filtering based enhanced medical image fusion using parameter-adaptive PCNN and local features-based fuzzy weighted matrices. Appl Intell 52, 7965–7984 (2022). https://doi.org/10.1007/s10489-021-02834-0.
https://www.kaggle.com/datasets/tarundalal/dangerous-insects-dataset
