Evaluating the integration of deep neural networks with machine learning classification algorithms to detect and classify eight pest species in tomato and wheat crops

Recognizing the challenges associated with conventional pest identification methods, which are labor-intensive, detrimental, and restrictive, the development of effective predictive models for accurate crop pest identification is paramount. This research utilized machine learning and deep learning techniques to extract features and classify tomato pests. Transfer learning was applied to a dataset of ۱۳۰۴ images of eight distinct tomato pests to expedite training. Critical pest characteristics were extracted employing convolutional neural network models, including AlexNet, GoogLeNet, ResNet۱۰۱, and VGG۱۶. Naïve Bayesian, Support Vector Machine, and K-Nearest Neighbors. With data augmentation and image preprocessing, ResNet۱۰۱ achieved the highest performance, attaining ۹۲.۴% accuracy and surpassing the other models. The classification capabilities of ResNet۱۰۱ and the SVM classifier were evaluated employing F۱ scores across eight pest species (Bemisia argentifolii, Helicoverpa armigera, Myzus persicae, Spodoptera exigua, Spodoptera litura, Thrips palmi, Tetranychus, and Zeugodacus cucurbitae), yielding F۱ scores of ۶۹/۷۴%, ۹۲/۸۷%, ۸۸/۸۸%, ۷۵/۶۹%, ۵۲/۶۵%, ۷۶/۹%, ۳۵/۳۵%, ۷۲%, and ۱۰۰%, respectively. The integration of machine learning and deep learning methodologies enables faster pest detection for experts and farmers, improves feature extraction, shortens training time, and ultimately increases crop yields and reduces economic losses.

کلیدواژه ها:

Tomato pest ، Machine Learning ، Deep Learning ، Feature Extraction ، transfer learning ، Grad-CAM