Color sorting is a technological operation performed with the aim of classifying compliant and noncompliant agricultural products in large-capacity industrial systems for agricultural product processing. This paper investigates the application of the YOLOv3 algorithm on raspberry images as a method developed for the detection, localization, and classification of objects based on convolutional neural networks (CNNs). To our knowledge, this is the first time a YOLO algorithm or CNN has been used with original images from the color sorter to focus on agricultural products. Results of the F1 measure were in the 92–97% range. Images in full resolution, 1024 × 1024, produced an average detection time of 0.37 s. The impact of the hyperparameters that define the YOLOv3 model as well as the impact of the application of the chosen augmentative methods on the model are evaluated. The successful classification of stalks, which is particularly challenging due to their shape, small dimensions, and variations, was achieved. The presented model demonstrates the ability to classify noncompliant products into four classes, some of which are appropriate for reprocessing. The software, including a graphic interface that enables the real-time testing of machine learning algorithm, is developed and presented.

Machine vision has been involved in the research and practical application of various phases of agricultural production since several decades ago. It includes classic image processing techniques that use different color spaces, as well as hyperspectral imaging, which can encompass near and far infrared, UV and Xrays. Apart from classic machine vision methods, the application of artificial neural networks and machine learning (especially deep learning) is currently causing considerable improvement in agricultural research and practice. In this paper, the principles of the selection of the different machine vision methods in different agricultural applications is considered. Special attention is paid to the examples of the applications of convolutional neural networks in sorting of agricultural products.

A promising concept of artificial intelligence that records intense developments in the field of digital imaging is Deep Learning (DL). More intensive research within this field has been recorded over the past two decades, and has been applied in the agricultural industry as well. This paper will describe the DL technology which represents a part of Machine Learning (ML), based on Convolutional Neural Networks (CNN). It takes a special application in a machine vision where it allows machines to learn from experience, adapt to new technologies, and perform human tasks. Input data can be from a variety of sources: from classic digital camera shots to satellite images, as well as from recordings obtained by means of hyperspectral, thermal and infrared cameras. The increasing popularity and use of trunks in agricultural fields is increasing, and the very application of these new technologies leads to the huge amount of data that needs to be processed in real time, therefore, DL algorithms are increasingly used. The paper will summarize the current and considered possible applications of CNN in primary and precise agriculture in the future.