Identification Of Butterfly Species Research Articles

Butterfly Image Classification Using Convulational Neural Network[CNN

Butterfly species identification through image classification is now a major use of computer vision, utilizing supervised learning methods to classify different types of butterflies based on images. This article gives a detailed overview of the latest progress in classifying butterfly images through supervised learning techniques on Google Colab, a widely-used cloud-based platform for machine learning projects. The review starts by discussing the significance of precise butterfly categorization for biodiversity research and conservation endeavors. It then goes into specifics about different methods used in supervised learning for this purpose, such as convolutional neural networks (CNNs), support vector machines (SVMs), and k-nearest neighbors (k-NN). The review discusses the pros and cons of using these methods on butterfly image data, emphasizing on accuracy, efficiency, and generalization. Special focus is placed on the preprocessing procedures necessary to improve image quality and extract features, including image augmentation, normalization, and feature scaling. The article also investigates various butterfly image datasets that are accessible to the public, analyzing how they are used for training and assessing classification models. Google Colab is highlighted as a potent instrument for creating and testing these models because of its convenience, user-friendliness, and compatibility with leading machine learning libraries such as TensorFlow and PyTorch. Furthermore, the article examines recent research and initiatives that have effectively utilized butterfly image categorization with Colab, demonstrating ideal methods and insight gained. Image Preprocessing, Feature Extraction, Machine Learning Libraries, TensorFlow, PyTorch, Image Augmentation, Publicly Available Datasets.

Diversity and colonization of butterfly (Lepidoptera) on Pasoso Island, Central Sulawesi

Various biotic and abiotic environmental factors influence the preservation of butterfly species in an area. The pressure and threat of extinction for butterflies that occupy relatively narrow and isolated habitats such as small islands are pretty large. This study aims to estimate the diversity of butterfly species, predict colonization, predict local extinction of butterflies, and conservation implications to minimize local extinction of butterflies in the habitat type on Pasoso Island, Central Sulawesi Province. Data was collected using field observations in four land cover types (primary forest, gardens, “ecoton” between garden-primary forests, and shrubs) on a 100 m line transect line. Each line transect is divided into point counts, with a distance between points of 100 m. The results of the identification of butterfly species in four habitat types on Pasoso Island found 16 species from 78 individual butterflies belonging to four families (Nymphalidae, Papilionidae, Pieridae, and Lycaenidae). Butterflies in all habitat types on Pasoso Island are in the moderate category. Species diversity in garden habitat types based on the analysis of the Shannon-Wiener index showed higher than other habitat types (H'=1.89), while the Ecoton habitat type has the lowest diversity value (H'=1.06). The results of the t-test showed that there was no significant difference in the diversity of butterfly species in the four habitats. The results showed that colonization and local extinction occurred in butterfly species on Pasoso Island. Four of the nine species are included in the extinct category, namely Apias albina, Family Lycaenidae, Ideopsis juventa, Phalanta alcippe celebensis Wallacea. Five other species were colonized, namely Arhopala araxes, Catochrysops strabo, Catopsilia pomona, Hypolimnas anomala, and Idea blanchardii.

Butterfly Species Identification from Natural Environment Based on Improved RetinaNet

Butterfly Species Identification from Natural Environment Based on Improved RetinaNet

Investigations of butterfly species identification from images in natural environments

It has been a challenging problem to identify species of butterflies, especially from images taken in natural environments. Therefore the First international butterfly species recognition competition was organized at the third Data Mining Competition in China in 2018, so as to find good solutions to this challenging problem. The baseline for the competition was based on the Faster R-CNN for it was the latest deep learning algorithm at that time. Nearly all the competition teams chose the Faster R-CNN, or its variations, to solve the problem. But the identification rates were not good enough, and Faster R-CNN is very time consuming. As a result we have been trying to find the most suitable algorithm to solve the butterfly species identification challenge. This paper will present some investigations we have undertaken in this field over the past two years, and show the results we have obtained. We propose a new partition and augmentation technique for the extremely unbalanced ecological butterfly database. We found that RetinaNet is, so far, the best deep learning algorithm to tackle butterfly species identification based on butterfly images taken in natural environments. The best result we obtained was 79.7% in terms of mAP (mean average precision). This is the best result compared to the state-of-the-art studies in this field on the same database so far.

Scientists by chance: reliability of non-structured primary biodiversity data. Insights from Italian Forums of Natural Sciences

Forums and social networks store a big deal of data on flora and fauna, collected especially by amateurs. To what extent are these data useful to contribute to biodiversity data systems? In this paper, we addressed the question about the "suitability for use" of primary biodiversity data by exploring two popular and valued Italian Forums of Natural Science (Forum Natura Mediterraneo and Forum Entomologi Italiani) and tried to assess their scientific potential. The aim of our work was to evaluate and discuss taxonomic reliability of the identification of butterfly species and the accuracy of their geographic locations. For each forum thread, we examined the posted images of butterflies, checked the diagnoses and georeferenced the observations from the textual descriptions provided by the users. Then, we compared each final identification by users with an independent identification by expert taxonomists. Looking at species level identifications, users identified 3764 out of 4029 specimens (93.4%) and experts agreed with them in 3649 cases: a high percentage agreement (po = 96.9%). As for the geographic data, we were able to georeferenced 97.9% of the observations (70% with an estimated extent less than 2500m). Results of this study, although limited to butterflies, suggest that the final identifications from forums show a surprisingly small bias and that the 'democratic' approach to taxonomy ultimately produces few uncertainties. The selected forums contain large amounts of primary biodiversity data in digital format, correctly identified and georeferenced with satisfactory accuracy and this capital is too valuable to remain unused. The formalization of collaborations with scientific projects and institutions would bring the forums in the area of “official” citizen science initiatives, giving the forums a role of citizens' scientific training. The recognition of a scientific role makes forum managers and users more deeply involved and data protection over time, currently entrusted to forum managers, would be greatly enhanced.

Deep learning techniques for automatic butterfly segmentation in ecological images

Automatic identification of butterfly species has attracted more and more attention due to the increasing demand for the accuracy and timeliness of butterfly species identification. Since the butterfly images we captured are usually ecological images, which not only have butterflies but also contain many irrelevant objects, such as leaves, flowers and other complex backgrounds. Therefore, segmenting butterflies from their ecological images is an issue that needs to be addressed prior to the tasks of identification and the segmentation quality directly affects the identification effect. However, the huge differences in butterflies, and the complexity of the natural environment make it very challenging to accurately segment butterflies from ecological images. Deep learning based methods are more promising for butterfly ecological image segmentation than traditional methods because they have powerful feature learning and representation ability. However, butterfly segmentation is still challenging when complex background interference occurs in images. To address this issue, we propose a dilated encoder network to capture more high-level features and get high-resolution output, which is both lightweight and accurate for automatic butterfly ecological image segmentation. In addition, we adopt the dice coefficient loss function to better balance the butterfly and non-butterfly regions. Experimental results on the public Leeds Butterfly dataset demonstrate that our method outperforms the state-of-the-art deep learning based image segmentation approaches.

Automatic identification for field butterflies by convolutional neural networks

In today’s competitive conditions, producing fast, inexpensive and reliable solutions are an objective for engineers. Development of artificial intelligence and the introduction of this technology to almost all areas have created a need to minimize the human factor by using artificial intelligence in the field of image processing, as well as to make a profit in terms of time and labor. In this paper, we propose an automated butterfly species identification model using deep neural networks. We collected 44,659 images of 104 different butterfly species taken with different positions of butterflies, the shooting angle, butterfly distance, occlusion and background complexity in the field in Turkey. Since many species have a few image samples we constructed a field-based dataset of 17,769 butterflies with 10 species. Convolutional Neural Networks (CNNs) were used for the identification of butterfly species. Comparison and evaluation of the experimental results obtained using three different network structures are conducted. Experimental results on 10 common butterfly species showed that our method successfully identified various butterfly species.

Automatic Identification of Butterfly Species Based on Gray-Level Co-occurrence Matrix Features of Image Block

In recent years, automatic identification of butterfly species arouses more and more attention in different areas. Because most of their larvae are pests, this research is not only meaningful for the popularization of science but also important to the agricultural production and the environment. Texture as a notable feature is widely used in digital image recognition technology; for describing the texture, an extremely effective method, graylevel co-occurrence matrix (GLCM), has been proposed and used in automatic identification systems. However, according to most of the existing works, GLCM is computed by the whole image, which likely misses some important features in local areas. To solve this problem, this paper presents a new method based on the GLCM features extruded from three image blocks, and a weight-based k-nearest neighbor (KNN) search algorithm used for classifier design. With this method, a butterfly classification system works on ten butterfly species which are hard to identify by shape features. The final identification accuracy is 98%.

Automatic identification of butterfly species based on HoMSC and GLCMoIB

Automatic identification is an efficient technology for the identification of butterfly species and pest control. As the major agriculture and forest pest, butterflies can be accurately classified based on the taxonomic characters. However, such identification can be only achieved by a few insect experts with years of experience. In the study, the shape and texture of butterflies were investigated for the automatic identification of butterfly species in the digital images: Histograms of multi-scale curvature (HoMSC) and gray-level co-occurrence matrix of image blocks (GLCMoIB) were used to describe the shape and texture of butterfly wings, respectively. To get an accurate identification result, a weight-based k-nearest neighbor classifier was designed. In addition, 750 images of 50 butterfly species were used for the identification test. The accuracy rate of this automatic identification method reached 98%. The result suggested that the HoMSC and GLCMoIB features can be efficient for the identification of butterfly species.

Handicraft of butterflies and moths (Insecta: Lepidoptera) in Bantimurung Nature Recreation Park and its implications on conservation

Putri IASLP. 2016 Handicraft of butterflies and moths (Insecta: Lepidoptera) in Bantimurung Nature Recreation Park and itsimplications on conservation. Biodiversitas 17: 823-831. The abundance of butterflies in Bantimurung Nature Recreation Park of Bantimurung-Bulusaraung National Park, South Sulawesi, Indonesia provides economic benefits to the community through butterfly’s handicrafts trading. This study aims to determine local species of commodified butterfly that are traded in various forms of craft and its implications for the conservation of butterflies. The study was conducted through the direct identification of butterfly species which are sold as crafts or deposited directly by the catchers to collectors. Data of commodified butterfly were collected using direct interviews.Data were analyzed by descriptive quantitative and qualitative. The results showed that there are 142 species of butterfly which are traded in the period of 2010-2015. The seller participants on butterfly handicrafts consist of the butterfly catchers, middlemen, craftsmen, stall employee, stall employers, and street vendors. The buyer participants consist of local tourists, tourists from outside district/province, traders from outside district/province, buyers from overseas and scientists or butterfly collectors. The butterfly price range was in between Rp. 500.00-Rp. 150,000.00/head at collectors’ level. The butterfly selling prices increased up to Rp. 7,500.00-Rp. 1,000,000.00 when they were processed into various souvenirs forms. Considering that there were so many traded butterfly souvenirs in the market, it raised an impression that there were more butterflies trapped for souvenir than free-living butterfly escaped from the trap. Commodification of butterflies needs to be regulated by setting the butterflies harvesting quota based on population in nature, sex, season and age (especially for female butterfly), accompanied by socializing rules of law, increasing public awareness about the importance of conservation butterflies, and creating new jobs for the people who depend on the butterflies trading.

Automatic identification of butterfly species based on local binary patterns and artificial neural network

Butterflies are classified firstly according to their outer morphological qualities. It is required to analyze genital characters of them when classification according to outer morphological qualities is not possible. Genital characteristics of a butterfly can be determined by using various chemical substances and methods. Currently, these processes are carried out manually by preparing genital slides of the collected butterfly through some certain processes. For some groups of butterflies molecular techniques should be applied for identification which is expensive to use. In this study, a computer vision method is proposed for automatically identifying butterfly species as an alternative to conventional identification methods. The method is based on local binary pattern (LBP) and artificial neural network (ANN). A total of 50 butterfly images of five species were used for evaluating the effectiveness of the proposed method. Experimental results demonstrated that the proposed method has achieved well recognition in terms of accuracy rates for butterfly species identification.

A vision system for automatic identification of butterfly species using a grey-level co-occurrence matrix and multinomial logistic regression

We present an application of image-processing techniques for identifying butterfly species as an alternative to conventional diagnostic methods. Grey-level cooccurrence matrix (GLCM) matrices are utilised to evaluate the surface texture features of butterflies' wings, which is an important character for identification. Eleven textural features were extracted from butterfly images and characterised by the texture average in four directions (0°, 45°, 90° and 135°) and distances (d = 1, 2, 3 and 4 pixels). We used 190 butterfly images belonging to 19 different species of the family Pieridae. The identification accuracy of the GLCM+MLR was 96.3% with tenfold cross validation. The methodology presented here classified the butterflies effectively. These findings suggest that the proposed MLR algorithm and GLCM texture features technique are feasible for the identification and classification of butterfly species.

Identification of butterfly based on their shapes when viewed from different angles using an artificial neural network

Identification of butterfly species is essential because they are directly associated with crop plants used for human and animal consumption. However, the widely used reliable methods for butterfly identification are not efficient due to complicated butterfly shapes. We previously developed a novel shape recognition method that uses branch length similarity (BLS) entropy, which is a simple branching network consisting of a single node and branches. The method has been successfully applied to recognize battle tanks and characterize human faces with different emotions. In the present study, we used the BLS entropy profile (an assemble of BLS entropies) as an input feature in a feed-forward back-propagation artificial neural network to identify butterfly species according to their shapes when viewed from different angles (for vertically adjustable angle, θ=±10°, ±20°, …, ±60° and for horizontally adjustable angle, φ=±10°, ±20°, …, ±60°). In the field, butterfly images are generally captured obliquely by camera due to butterfly alignment and viewer positioning, which generates various shapes for a given specimen. To generate different shapes of a butterfly when viewed from different angles, we projected the shapes captured from top-view to a plane rotated through angles θ and φ. Projected shapes with differing θ and φ values were used as training data for the neural network and other shapes were used as test data. Experimental results showed that our method successfully identified various butterfly shapes. In addition, we briefly discuss extension of the method to identify more complicated images of different butterfly species.

A Computer Vision System for the Automatic Identification of Butterfly Species via Gabor-Filter-Based Texture Features and Extreme Learning Machine: GF+ELM

Butterflies are classified first according to their outer morphological qualities. It is required to analyze their genital characters when classification according to their outer morphological qualities is not possible. The genital characters of butterflies can be obtained using various chemical substances and methods; however, these processes can only be carried out with some certain expenses. Furthermore, the preparation of genital slides is time-consuming since it requires specific processes. In this study, a new method based on the extreme learning machine (ELM) and Gabor filters (GFs), which is an image processing technique, was used for the identification of butterfly species as an alternative to conventional diagnostic methods. GFs have been recognized as a very useful tool in texture analysis, due to their optimal localization properties in both the spatial and frequency domains, and have been found to have widespread use in computer vision applications. To obtain the appropriate features from butterfly images in the spatial domain, 20 filters were designed for the various angles and frequencies (5 frequencies and 4 orientations). The diagnosing of butterflies was performed through ELM, with texture features based on GFs. The classification process was performed with a 75%–25% training-test set for different activation functions and the recognition performance value was obtained as 97.00%. In addition, the recognition success rates with ELM were compared to other machine learning methods and it was seen that ELM has a more significant success rate in butterfly identification than other methods. As a result, the proposed method is a suitable machine vision system for detecting butterfly species

Application of artificial neural network for automatic detection of butterfly species using color and texture features

Butterflies can be classified by their outer morphological qualities, genital characteristics that can be obtained using various chemical substances and methods which are carried out manually by preparing genital slides through some certain processes or molecular techniques which is a very expensive method. In this study, a new method which is based on artificial neural networks (ANN) and an image processing technique was used for identification of butterfly species as an alternative to conventional diagnostic methods. Five texture and three color features obtained from 140 butterfly images were used for identification of species. Texture features were obtained by using the average of gray level co-occurrence matrix (GLCM) with different angles and distances. The accuracy of the purposed butterfly classification method has reached 92.85Â %. These findings suggested that the texture and color features can be useful for identification of butterfly species.

Butterfly species identification by branch length similarity entropy

We previously developed a shape recognition methodology that uses “branch length similarity” (BLS) entropy, which is defined as a simple branching network consisting of a single node and branches. The simple network is referred to as a “unit branching network” (UBN). Our approach involves obtaining BLS entropy profiles from UBNs created by joining each pixel in the outline of a shape with every other pixel in the shape's border. The profiles successfully characterize the shapes by comparing their BLS entropy profiles. Presently, we modified this approach to facilitate its application to butterfly species identification by partitioning and weighting the entropy profile. As a test, we identified the butterfly species Colias erate, Parnassius bremeri, Eurema hecabe, Gonepteryx rhamni, and Papilio maackii. Each species group consisted of 10 specimens. We used wing shape to identify a species. We extracted evenly spaced x–y coordinates of boundary pixels for the wing shapes in a counter counterclockwise direction. The number of the pixels was 749. We then sequentially partitioned 749 x–y pairs into 15 groups, calculated entropy profiles for the groups, and weighted the profiles. The profiles were combined in order, resulting in a single weighted BLS entropy profile for a wing's shape. Subsequently, we statistically compared the correlation coefficient among the weighted BLS profiles. Our experimental results showed that this method was statistically successful for butterfly species identification. The advantage of the partitioning and weighting process in shape recognition is also discussed.

Identification of butterfly species with a single neural network system

Growing interest in conservation and biodiversity increased the demand for accurate and consistent identification of biological objects, such as insects, at the level of individual or species. Among the identification issues, butterfly identification at the species level has been strongly addressed because it is directly connected to the crop plants for human food and animal feed products. However, so far, the widely-used reliable methods were not suggested due to the complicated butterfly shape. In the present study, we propose a novel approach based on a back-propagation neural network to identify butterfly species. The neural network system was designed as a multi-class pattern classifier to identify seven different species. We used branch length similarity (BLS) entropies calculated from the boundary pixels of a butterfly shape as the input feature to the neural network. We verified the accuracy and efficiency of our method by comparing its performance to that of another single neural network system in which the binary values (0 or 1) of all pixels on an image shape are used as a feature vector. Experimental results showed that our method outperforms the binary image network in both accuracy and efficiency.