Akademik Veri Yönetim Sistemi
1st International Computer Science, Engineering and Information Technology Congress (ICSITY 2022), Warszawa, Polonya, 29 - 30 Eylül 2022, ss.63-70
Deep learning method is a method that provides superior success in detecting and classifying objects in an image. In this method, an algorithm is generally used that distinguishes objects in complex images. Convolutional neural networks (CNN), especially used in deep learning, are an important research topic in computer-aided recognition and are used in many areas. In recent years, recognition and classification have been made in the agricultural area, as in many areas. Especially with the introduction of special robotic applications into the agricultural area, it is ensured that the workforce and loss in agricultural activities are reduced. Deep learning methods used in advanced robotic applications help to collect objects efficiently by distinguishing them from each other.
In this study, a vegetable recognition and classification system are proposed to support special robotic systems used in agriculture. In this application, the MobileNet model structure, which was previously accepted in the literature, was used. In this model, recognition and classification were made using a data set containing 4 different types of vegetables. According to the evaluation results, it was seen that the vegetable types in the data set were classified correctly with a high success rate.