Convolutional neural network architecture for plant seedling classification

Abstract

Weed control is a challenging problem that may face crops productivity. Weeds are perceived as an important problem because they conduce to reduce crop yields due to the expanding competition for nutrients, water, and sunlight besides they serve as hosts for diseases and pests. Thus, it is crucial to identify weeds in early growth in order to avoid their side effects on crops growth. Previous conventional machine learning technologies exploited for discriminating crops and weeding species faced challenges of effectiveness and reliability of weed detection at preliminary stages of growth. This work proposes the application of deep learning technique for plant seedling classification. A new Convolutional Neural Networks (CNN) architecture is designed to classify plant seedlings at their early growth stages. The presented technique is appraised using plant seedlings dataset. Average accuracy, precision, recall, and F1-score are utilized as evaluation metrics. The results reveal the capability of the proposed technique in discriminating among 12 species (3 crops and 9 weeds). The system achieved 94.38% average classification accuracy. The proposed system is compared with existing plant seedling systems. The results demonstrate that the proposed method outperforms the existing methods.