Automatic recognition of pollen bearing bees can provide important information both for pollination monitoring and for assessing the health and strength of bee colonies, with the consequent impact on people's lives, due to the role of bees in the pollination of many plant species. In this paper, we analyse some of the Convolutional Neural Networks (CNN) methods for detection of pollen bearing bees in images obtained at hive entrance. In order to show the influence of colour we preprocessed the dataset images. Studying the results of nine state-of-the-art CNNs, we provide a baseline for pollen bearing bees recognition based in deep learning. For some CNNs the best results were achieved with the original images. However, our experiments showed evidence that DarkNet53 and VGG16 have superior performance against the other CNNs tested, with unsharp masking preprocessed images, achieving accuracy results of 99.1% and 98.6%, respectively.

No abstract

material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.