A Hybrid Grey Wolves Optimizer and Convolutional Neural Network for Pollen Grain Recognition

A Hybrid Grey Wolves Optimizer and Convolutional Neural Network for Pollen Grain Recognition

Hanane Menad (EEDIS Laboratory, Djillali Liabes University, Sidi Bel Abbes, Algeria), Farah Ben-naoum (EEDIS Laboratory, University of Djillali Liabes, Sidi Bel Abbes, Algeria) and Abdelmalek Amine (GeCoDe Laboratory, Department of Computer Science, Tahar Moulay University of Saida, Algeria)
Copyright: © 2020 |Pages: 23
DOI: 10.4018/IJSIR.2020070104

Abstract

Melissopalynology, or pollen analysis of honey, is one of the areas that benefited greatly from image processing and analysis techniques, where melissopalynology is the science that studies the pollen contained in honey, using a microscopic examination. Nowadays, developing an automatic classification system for pollen identification presents a challenge. This article presents a metaheuristic for image segmentation to detect pollen grains in images. It is a swarm intelligence technique inspired from grey wolf hunting behavior in nature, centered around respecting the hierarchy of a pack. It was tested on a set of microscopic images of pollen grains. To evaluate pollen detection, we represented the detected pollen grains using two methods, grey-level based representations where we kept grey value of each pixel, and a binary mask-based technique, where a pixel could have only two values (1 or 0). Then, we used a convolutional neural network (CNN) technique for image classification to predict the specie of each pollen. The proposed system was tested on a set of microscopic images of pollen grains. The obtained performance measures of the system proved that the system is very successful.
Article Preview
Top

Introduction

Honey is a very complex natural product that contains sugars, organic acids, amino acids, proteins, minerals, lipids, aroma compounds, flavonoids, vitamins, pigments, waxes, pollen grains, enzymes and other phytochemicals (Gomes, 2010). In light of the diversity of the flower mellific, there are various honeys which distinguish themselves by their composition, directly dependent on the origin of the nectar and on the miellat, climates, the conditions environmental and the skill of the beekeepers (Nair, 2014). Honey pollen profile reflects forest vegetation, floral diversity and species composition of the plants foraged by honeybees. The relative pollen frequency is used to precise the main origin of the nectar or reflects the geographical originality, also the good knowledge of honey types constitutes the essential basis of a rational marketing, being also used as a traceability tool by food control institutions (Bryant, 2001) (Corvucci, 2015). Honeys can be classified according to the origin of the pollen grain as monofloral if it is dominated from one particular plant, otherwise if the honey is given by several vegetable species it may be classified as polyfloral (Moussa, 2015). And in order to avoid falsification, the international honey commission has created standards, which are used to determine the quality of the honey (Bogdanov, 2002). The analysis of honey pollen or melissopalynology has a big importance for the control of honey quality (Popek, 2002; Svečnjak, 2015; Ponnuchamy, 2014; Brownlee, 2011).

Nowadays, image processing is one of most active research fields, with the success seen in this area, it has been applied in many other research areas, because it has proved that their interest in areas such as image restoration or three-dimensional vision, and it is being justified in other previously unexplored areas such as diagnostic assistance. Moreover, the image represents one of the richest sources of information. Because of the variety of possibilities that this information offers, in a myriad of areas, there has been a great deal of enthusiasm for research in the field of computer vision, especially since the advent of digital images. In this regard, Numerous researches has been carried out in this field, and remains one of the most studied areas. There are basically two levels of automatic image processing. The first is a low-level treatment dedicated to acquisition, compression, segmentation, improvement or restoration. The second level is a high-level processing dedicated to symbolic image analysis operations, such as description, recognition or interpretation, in order to extract information (Benaichouche, 2014).

Figure 1 shows the manual pollen recognition process done by biologists, in which the biologist must detect in the image the pollen which equals to make the segmentation of the image, and is based on the information observed on this pollen namely: the morphology or the form, the pores and grooves whose number and the provision differs from one species to another. The difficulty is that many botanical species has a similar floral morphology which makes the processes of the pollen recognition very difficult for the human.

Figure 1.

Manual pollen recognition process done by biologists

IJSIR.2020070104.f01

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 12: 4 Issues (2021): Forthcoming, Available for Pre-Order
Volume 11: 4 Issues (2020)
Volume 10: 4 Issues (2019)
Volume 9: 4 Issues (2018)
Volume 8: 4 Issues (2017)
Volume 7: 4 Issues (2016)
Volume 6: 4 Issues (2015)
Volume 5: 4 Issues (2014)
Volume 4: 4 Issues (2013)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing