基于颜色特征的昆虫自动鉴定方法

对昆虫实现自动鉴定是将相关专家从大量重复性的鉴定工作中解脱出来的最有效的方法之一。本章文着重论述昆虫自动鉴定系统鉴定的基本原理,开发步骤和使用昆虫自动鉴定系统进行自动鉴定的步骤,基于颜色特征的昆虫自动鉴定系统开发中在图像获取、颜色特征提取方面的难点和相应的解决方案。最后,经过测试开发出的基于颜色特征的蝴蝶自动鉴定系统在对43种蝴蝶的自动鉴定上取得了95.2%准确率。     

Traditional ribotyping shows a higher discrimination than the automated RiboPrinter system in typing Vibrio cholerae O1

Sixteen clinical Vibrio cholerae O1 strains from four different countries were selected for comparison by traditional ribotyping and an automated RiboPrinter system for identification and discrimination purposes. Automated ribotyping, which routinely uses the restriction enzyme EcoRI for typing all bacterial species, produced only five different ribotypes compared with 10 different EcoRI ribotypes obtained by the traditional method. Traditional and automated ribotyping using the restriction enzyme BglI, which is recommended for the ribotyping of V. cholerae, produced 10 and seven different ribotypes, respectively. The lower discrimination shown by the RiboPrinter system was caused mainly by an inability to differentiate closely located fragments due to a lower resolution and electrophoresis conditions, a parameter which cannot be changed in the automated system. The RiboPrinter system includes a database for bacterial identification. However, none of the V. cholerae O1 strains studied showed EcoRI ribotype patterns which matched any of the patterns included in the database. In conclusion, the existing RiboPrinter system is not adequate for taxonomic identification and classification of V. cholerae O1.     

实蝇科果实蝇属昆虫数字图像自动识别系统的构建和测试

针对双翅目实蝇科果实蝇属昆虫的自动识别,本文提出利用翅及中胸背板图像的局部二进制模式（local binary pattern, LBP）特征,采用Adaboost算法, 设计和开发“实蝇科果实蝇属昆虫数字图像自动识别系统”（Automated Fruit fly Identification System-Bactrocera, AFIS-B）。该系统包括图像采集、图像裁剪、预处理、特征提取、分类器设计、识别和显示,共7个模块。研究结果表明： LBP特征可以有效鉴别实蝇科果实蝇属昆虫;在对实蝇科果实蝇属8个种的测试中, 该系统表现出较高的准确性和稳定性,平均识别率可达80%以上。此外,还对果实蝇属昆虫翅膀及中胸背板图像在光照不均匀、姿态扭曲、样本受损及样本量大小等不同条件下的识别率进行了试验测试。结果表明, 该系统对测试样本的光照不均匀、 姿态扭曲和样本受损都表现出良好的鲁棒性,正确识别率与训练集样本各个种数量在一定条件下明显正相关,与训练集样本物种总量负相关。该项研究为实蝇科有害昆虫自动识别系统的构建及实际应用提供了理论、 方法及基础数据的支撑, 亦可为其他昆虫自动识别系统的研究和构建提供有益借鉴。 关键词：     

Automated identification of live moths (Macrolepidoptera) using digital automated identification System (DAISY)

Two hundred and thirty‐seven species of Macrolepidoptera were light trapped at Treborth Botanical Garden, Gwynedd, UK. Live adults were digitally imaged using a simple, inexpensive method suitable for field use, then released. Inconsistent lighting, variation in resting posture and inclusion of worn individuals produced image sets high in intraspecific variation. Thirty‐five common species were selected to provide training images for the Digital Automated Identification SYstem (DAISY). Twenty individuals per species were pre‐processed to standardize size and posture and to enhance features. The right forewing of each was highlighted manually and the pattern rendered polar and greyscale for DAISY analysis. Despite poor quality of some images, 83% of unknown species were identified correctly. The best species had 100% correct identification and the worst 35%. The most poorly identified images were those of moths that had lost scales or been unevenly illuminated. The precision with which the forewing was highlighted affected performance. When highlighted carefully, Laothoe populi was identified correctly twice as successfully as when the same image was highlighted poorly. Size of the training set was also important. Sets of 5, 10, 15 and 20 training images, plotted against performance produced a curve of diminishing returns. Colourimages and inclusion of size should improve accuracy.     

Automatic identification of large collections of protein-coding or rRNA sequences

The number of available genomic sequences is growing very fast, due to the development of massive sequencing techniques. Sequence identification is needed and contributes to the assessment of gene and species evolutionary relationships. Automated bioinformatics tools are thus necessary to carry out these identification operations in an accurate and fast way. We developed HoSeqI (Homologous Sequence Identification), a software environment allowing this kind of automated sequence identification using homologous gene family databases. HoSeqI is accessible through a Web interface (http://pbil.univ-lyon1.fr/software/HoSeqI/) allowing to identify one or several sequences and to visualize resulting alignments and phylogenetic trees. We also implemented another application, MultiHoSeqI, to quickly add a large set of sequences to a family database in order to identify them, to update the database, or to help automatic genome annotation. Lately, we developed an application, ChiSeqI (Chimeric Sequence Identification), to automate the processes of identification of bacterial 16S ribosomal RNA sequences and of detection of chimeric sequences.     

Sexual size dimorphism and sex identification using morphological traits of two Aegithalidae species

The black-throated tit, Aegithalos concinnus, and long-tailed tit, A. caudatus, are two widely-distributed species of Aegithalidae. They are thought to be monomorphic and thus difficult to differentiate between sexes in the field. We determined the sex of 296 black-throated tits and 129 long-tailed tits using DNA analysis, evaluated their sexual size dimorphism, and developed discriminant models to identify sex based on morphometries, including bill length, bill depth, bill-head length, maximum tarsus length, tarsus width, wing length, tail length, total body length, and body mass. Both species were sexually dimorphic in size, with males having larger measurements than females, except for bill length in black-throated tits, and both bill length and body mass in long-tailed tits. Moreover, both species showed similar sexual size dimorphism (SSD) among the morphological features, with tail length having the highest SSD value. The multivariate discriminant models for sex identification had an accuracy of 82% for both species, which was slightly higher than the best univariate discriminatory model for each species. Because of the complicated nature of the multivariate model, we recommend univariate models for sex identification using D = 0.491 × tail length - 24.498 (accuracy 80%) for black-throated tits and D = 0.807 × wing length - 45.934 (accuracy 78%) for long-tailed tits. Females in both species showed generally higher morphological variation than did males, resulting in lower accuracies in all discriminate functions regardless of univariate or multivariate approach. This could be the result of a sex-biased selective pressure in which males have higher selective pressures for these morphological features.     

昆虫图像自动鉴别技术

昆虫是地球上物种多样性最为丰富的生物类群,其物种鉴定任务复杂而艰巨,可靠的物种鉴定是开展昆虫学工作的重要基础之一。当前,国内外的人工昆虫物种鉴定能力均不能满足实际需求,因而人们开始不断探索利用计算机自动鉴定昆虫的原理和方法。目前,模式识别技术的迅猛发展已为昆虫图像的自动鉴定提供可能。文章概述昆虫图像自动鉴定技术研究的历史与现状,总结主要原理和方法,介绍工作流程,并展望发展前景。     

Manta Matcher: automated photographic identification of manta rays using keypoint features

For species which bear unique markings, such as natural spot patterning, field work has become increasingly more reliant on visual identification to recognize and catalog particular specimens or to monitor individuals within populations. While many species of interest exhibit characteristic markings that in principle allow individuals to be identified from photographs, scientists are often faced with the task of matching observations against databases of hundreds or thousands of images. We present a novel technique for automated identification of manta rays (Manta alfredi and Manta birostris) by means of a pattern‐matching algorithm applied to images of their ventral surface area. Automated visual identification has recently been developed for several species. However, such methods are typically limited to animals that can be photographed above water, or whose markings exhibit high contrast and appear in regular constellations. While manta rays bear natural patterning across their ventral surface, these patterns vary greatly in their size, shape, contrast, and spatial distribution. Our method is the first to have proven successful at achieving high matching accuracies on a large corpus of manta ray images taken under challenging underwater conditions. Our method is based on automated extraction and matching of keypoint features using the Scale‐Invariant Feature Transform (SIFT) algorithm. In order to cope with the considerable variation in quality of underwater photographs, we also incorporate preprocessing and image enhancement steps. Furthermore, we use a novel pattern‐matching approach that results in better accuracy than the standard SIFT approach and other alternative methods. We present quantitative evaluation results on a data set of 720 images of manta rays taken under widely different conditions. We describe a novel automated pattern representation and matching method that can be used to identify individual manta rays from photographs. The method has been incorporated into a website (mantamatcher.org) which will serve as a global resource for ecological and conservation research. It will allow researchers to manage and track sightings data to establish important life‐history parameters as well as determine other ecological data such as abundance, range, movement patterns, and structure of manta ray populations across the world.     

Isolation and automated ribotyping of Mycobacterium lentiflavum from drinking water distribution system and clinical specimens

Automated ribotyping as a tool for identifying of nontuberculous mycobacteria was evaluated. We created a database comprising of riboprints of 60 strains, representing 32 species of nontuberculous mycobacteria. It was shown that combined ribopatterns generated after digestion with EcoRI and PvuII were distinguishable between species of both slow-growing and rapid-growing mycobacteria. The findings were in good agreement with the 16S rRNA gene sequencing results, allowing correct identification of Mycobacterium lentiflavum isolated from clinical specimens and from biofilms growing in public water distribution system. The automated ribotyping was powerful in discriminating between M. lentiflavum and closely related species M. simiae and M. palustre. Mycobacterium lentiflavum strains from drinking water biofilms were resistant to two to four antimycobacterial drugs. The drinking water distribution system may, thus, be a source of nontuberculous mycobacteria resistant to multiple drugs.     

Interactive and automated application of virtual microscopy

Virtual microscopy can be applied in an interactive and an automated manner. Interactive application is performed in close association to conventional microscopy. It includes image standardization suitable to the performance of an individual pathologist such as image colorization, white color balance, or individual adjusted brightness. The steering commands have to include selection of wanted magnification, easy navigation, notification, and simple measurements (distances, areas). The display of the histological image should be adjusted to the physical limits of the human eye, which are determined by a view angle of approximately 35 seconds. A more sophisticated performance should include acoustic commands that replace the corresponding visual commands. Automated virtual microscopy includes so-called microscopy assistants which can be defined similar to the developed assistants in computer based editing systems (Microsoft Word, etc.). These include an automated image standardization and correction algorithms that excludes images of poor quality (for example uni-colored or out-of-focus images), an automated selection of the most appropriate field of view, an automated selection of the best magnification, and finally proposals of the most probable diagnosis. A quality control of the final diagnosis, and feedback to the laboratory determine the proposed system. The already developed tools of such a system are described in detail, as well as the results of first trials. In order to enhance the speed of such a system, and to allow further user-independent development a distributed implementation probably based upon Grid technology seems to be appropriate. The advantages of such a system as well as the present pathology environment and its expectations will be discussed in detail.     

Artificial intelligence in pest insect monitoring

Abstract Global problems of hunger and malnutrition induced us to introduce a new tool for semi‐automated pest insect identification and monitoring: an artificial neural network system. Multilayer perceptrons, an artificial intelligence method, seem to be efficient for this purpose. We evaluated 101 European economically important thrips (Thysanoptera) species: extrapolation of the verification test data indicated 95% reliability at least for some taxa analysed. Mainly quantitative morphometric characters, such as head, clavus, wing, ovipositor length and width, formed the input variable computation set in a Trajan neural network simulator. The technique may be combined with digital image analysis.     

桔小实蝇幼体及成虫残体DNA条形码识别技术的建立与应用

实蝇类害虫多为国内外检疫对象, 其鉴定识别方法主要依据成虫的外部形态特征, 而传统的形态学识别法对口岸经常截获的幼体及残缺的虫体, 则无能为力。本研究以桔小实蝇Bactrocera dorsalis的幼体（卵、 幼虫、 蛹）以及成虫残体（足、 翅、 头部、 胸部、 腹部）为对象, 利用 DNA 条形码技术, 构建实蝇类害虫快速鉴定技术体系, 并以其他4种常见实蝇（包括番石榴实蝇B. correcta、 瓜实蝇B. cucurbitae、 南亚果实蝇B. tau、 柑桔大实蝇B. minax）为对象对该技术体系进行应用验证。结果显示, 桔小实蝇幼体以及成虫残体的碱基序列与数据库中靶标种COⅠ基因碱基序列的一致性为99.51%~99.84%, 其他4种实蝇相应序列与数据库中靶标种COⅠ基因序列的一致性分别为100%, 100%, 99.81%~99.83%和100%; 以邻接法（NJ法）构建系统发育树, 靶标种实蝇均与数据库中对应种实蝇聚为一支, 且置信度均为100%。以K2-P模型计算种内及种间遗传距离得出, 5种实蝇的种间遗传距离为0.0597~0.2363, 平均为0.1693; 种内遗传距离为0.0000~0.0041, 平均为0.0019。这些结果表明, 基于DNA条形码的物种识别技术完全可用于口岸截获的实蝇类害虫幼体及残体的准确鉴定。     

Towards automation of glycomic profiling of complex biological materials

Glycosylation is considered one of the most complex and structurally diverse post-translational modifications of proteins. Glycans play important roles in many biological processes such as protein folding, regulation of protein stability, solubility and serum half-life. One of the ways to study glycosylation is systematic structural characterizations of protein glycosylation utilizing glycomics methodology based around mass spectrometry (MS). The most prevalent bottleneck stages for glycomic analyses is laborious sample preparation steps. Therefore, in this study, we aim to improve sample preparations by automation. We recently demonstrated the successful application of an automated high-throughput (HT), glycan permethylation protocol based on 96-well microplates, in the analysis of purified glycoproteins. Therefore, we wanted to test if these developed HT methodologies could be applied to more complex biological starting materials. Our automated 96-well-plate based permethylation method showed very comparable results with established glycomic methodology. Very similar glycomic profiles were obtained for complex glycoprotein/protein mixtures derived from heterogeneous mouse tissues. Automated N-glycan release, enrichment and automated permethylation of samples proved to be convenient, robust and reliable. Therefore we conclude that these automated procedures are a step forward towards the development of a fully automated, fast and reliable glycomic profiling system for analysis of complex biological materials.     

Citizen monitoring of invasive species: wing morphometry as a tool for detection of alien Tetropium species

The increasing threat of alien wood‐boring insect has resulted in the initiation of large‐scale monitoring programmes. These programmes are most often based on pheromone‐bailed traps, which allow the early detection and monitoring of invasive species. This approach is expensive because it entails the processing and accurate identification of large numbers of specimens. One of the most often suggested solutions to this problem is citizen participation in the monitoring of invasive species. Such an approach has the potential for reducing costs as well as providing data from a larger number of sites. However, citizens vary in taxonomic expertise and experience which can result in identification errors. This may be particularly important in the case of wood borers which include many morphologically similar species. In this study, we develop and discuss a semi‐automated method of identifying four morphologically similar and invasive Tetropium spp. wood borers as a potential tool for citizen‐based monitoring programmes. Identification is based on wing measurements and requires neither specialist knowledge nor expensive equipment. The method correctly identified the species of Tetropium with an error ranging from 1.3% for T. fuscum to 7.5% for T. cinnamopterum. We found that experience level of the individual user was not essential for correct identification; on average, inexperienced volunteers correctly identified the Tetropium species in 93% of cases. Further development of this method may be a significant step to overcoming the taxonomical impediment to citizen monitoring of taxonomically challenging groups of insects.     

Vegetation mapping using hierarchical object‐based image analysis applied to aerial imagery and lidar data

Aims: The primary objective of this study is to map the distribution and quantify the cover of vegetation alliances over the entirety of San Clemente Island (SCI). To this end, we develop and evaluate the mapping method of hierarchical object‐based classification with a rule‐based expert system. Location: San Clemente Island, California, USA. Methods: We developed and tested an approach based on hierarchical object‐based classification with a rule‐based expert system to effectively map vegetation communities on SCI following the Manual of California Vegetation classification system. In this mapping approach, the shrub species defining each vegetation community and non‐shrub growth forms were first mapped using aerial imagery and lidar data, then used as input in an automated mapping rule set that incorporates the percent cover rules of a field‐based mapping rule set. Results: The final vegetation map portrays the distribution of 19 vegetation communities across SCI, with the largest areas comprised of California Annual and Perennial Grassland (35%) and three types of coastal sage scrub and maritime succulent scrub, comprising a combined 53% of the area. Map accuracy was assessed to be 79% based on fuzzy methods and 61% with a traditional accuracy assessment. The accuracy of tree identification was assessed to be 81%, but species‐level tree accuracy was 45%. Conclusions: Semi‐automated approaches to vegetation community mapping can produce repeatable maps over large spatial extents that facilitate ecological management efforts. However, some low‐statured shrub community types were difficult to differentiate due to patchy canopies of co‐occurring species including abundant non‐native grasses characteristic of complex disturbance histories. Species‐level tree mapping accuracy was low due to the difficulty of identifying species within poorly illuminated canyons, resulting from sub‐optimal image acquisition timing.     

Molecular characterisation and development of rapid molecular methods to identify species of Gracilariaceae from the Atlantic coast of Morocco

In Gracilariaceae, species identification is traditionally based on gross morphology; therefore the taxonomic status of terete individuals remains frequently problematic due to the lack of diagnostic characters to identify specimens. Different morphospecies have been recorded along the Atlantic coast of Morocco; however, no clear diagnostic characters were available to discriminate between terete species. Rapid molecular techniques have been developed recently to resolve many taxonomic problems and to re-assess the global diversity and biogeography in algae. In this study, molecular markers were used as DNA barcoding to characterise species. The sequence of the Rubisco spacer allowed identification of six species of Gracilariaceae: Gracilaria gracilis, Gracilaria dura, Gracilaria conferta, Gracilaria vermiculophylla, Gracilaria multipartita and Gracilariopsis longissima. In order to identify species with certainty, two simple and rapid methods based on the amplification of rDNA ITS and PCR-RFLP of the large subunit of the Rubisco were developed.     

A ROLE FOR LEARNING IN POPULATION DIVERGENCE OF MATE PREFERENCES

Learning and other forms of phenotypic plasticity have been suggested to enhance population divergence. Mate preferences can develop by learning, and species recognition might not be entirely genetic. We present data on female mate preferences of the banded demoiselle (Calopteryx splendens) that suggest a role for learning in population divergence and species recognition. Populations of this species are either allopatric or sympatric with a phenotypically similar congener (C. virgo). These two species differ mainly in the amount of wing melanization in males, and wing patches thus mediate sexual isolation. In sympatry, sexually experienced females discriminate against large melanin wing patches in heterospecific males. In contrast, in allopatric populations within the same geographic region, females show positive (“open‐ended”) preferences for such large wing patches. Virgin C. splendens females do not discriminate against heterospecific males. Moreover, physical exposure experiments of such virgin females to con‐ or hetero‐specific males significantly influences their subsequent mate preferences. Species recognition is thus not entirely genetic and it is partly influenced by interactions with mates. Learning causes pronounced population divergence in mate preferences between these weakly genetically differentiated populations, and results in a highly divergent pattern of species recognition at a small geographic scale.     

The utility of wing morphometrics for assigning type specimens to cryptic bumblebee species

Since the beginning of taxonomy, species have been described based on morphology, but the advent of using semio-chemicals and genetics has led to the discovery of cryptic species (i.e. morphologically similar species). When a new cryptic species is described, earlier type specimens have to be re-evaluated, although this process can be challenging as only nondestructive methods ought to be used in order to preserve the integrity of the type specimens. Methods should allow comparison with recently collected specimens clustered based on chemical, ethological and/or genetic traits with old specimens (i.e. type specimens) where only morphological traits are available. Here we develop a method based on geometric morphometric analyses of wing shape for a taxonomically challenging group of bumblebees, the subgenus Alpinobombus Skorikov. We consider nine monophyletic taxa (including several cryptic species) to assess the accuracy of this method to discriminate the taxa based on their wing shape and then to attribute type specimens using a leave-one-out cross-validation procedure. We show that, for these bees, wing shape is taxon-specific, except for two sister taxa for which the species status is still debated. Moreover, for most of the taxa, type specimens were correctly attributed with high posterior probabilities of attribution, except for a few type specimens corresponding to the same two sister taxa where taxa delimitation based on wing shape was previously the subject of discussion. Our study highlights the potential of geometric morphometric analyses to help in the re-attribution of type specimens when the existence of cryptic species is revealed.     

Development and Validation of a Wing Key to Improve Harvest Management of Alcids in the Northwest Atlantic

Abstract: Murres (thick-billed [Uria lomvia] and common [U. aalge]) are legally hunted along the coast of Newfoundland and Labrador, Canada. Razorbills (Alca torda) are also incidentally taken. Only irregular estimates of the total murre harvest are available, so a tool to derive estimates of age- and species-specific harvest is required to effectively monitor the hunt and manage a sustainable harvest. We collected 293 murre and razorbill wings from hunters between 1999–2004, with the goal of identifying wing characteristics that could be used to discriminate age and species. We found that murres and razorbills could be reliably aged (first-yr vs. older) on the basis of molt limits of greater wing coverts. Using a discriminant function (DF) incorporating length of the first primary and second secondary feather, we classified 95–96% of common murres and 99–100% of thick-billed murres correctly to species. First-year thick-billed and common murres also differed in number of pale secondary coverts (median = 12 and 3, respectively), providing another species-specific trait. We developed a key to age and assign species based on these results. We assessed applicability and accuracy of the wing-key with novice observers, who differentiated between murre and razorbill wings using feather-pattern coloration with high accuracy (95 ± 9%) and were able to differentiate between the 2 murres species using 3 techniques: visual assessment of wing shape (83 ± 14% accuracy), the DF (94 ± 6%), and number of worn secondary coverts for first-year birds only (83 ± 5%). Experience increased success rates of aging and species classification using wing shape and number of worn secondary coverts but not using the DF. Despite differences in measurement accuracy and repeatability among observers, the DF proved to be robust. Our results will facilitate implementation of a species composition survey for the murre hunt and will improve identification rates of carcasses found during beached bird surveys in the Northwest Atlantic, aiding in monitoring of alcid populations vulnerable to anthropogenic activities.     

Wing shape as a potential discriminator of morphologically similar pest taxa within the Bactrocera dorsalis species complex (Diptera: Tephritidae)

Four morphologically cryptic species of the Bactrocera dorsalis fruit fly complex (B. dorsalis s.s., B. papayae, B. carambolae and B. philippinensis) are serious agricultural pests. As they are difficult to diagnose using traditional taxonomic techniques, we examined the potential for geometric morphometric analysis of wing size and shape to discriminate between them. Fifteen wing landmarks generated size and shape data for 245 specimens for subsequent comparisons among three geographically distinct samples of each species. Intraspecific wing size was significantly different within samples of B. carambolae and B. dorsalis s.s. but not within samples of B. papayae or B. philippinensis. Although B. papayae had the smallest wings (average centroid size=6.002 mm±0.061 SE) and B. dorsalis s.s. the largest (6.349 mm±0.066 SE), interspecific wing size comparisons were generally non-informative and incapable of discriminating species. Contrary to the wing size data, canonical variate analysis based on wing shape data discriminated all species with a relatively high degree of accuracy; individuals were correctly reassigned to their respective species on average 93.27% of the time. A single sample group of B. carambolae from locality 'TN Malaysia' was the only sample to be considerably different from its conspecific groups with regards to both wing size and wing shape. This sample was subsequently deemed to have been originally misidentified and likely represents an undescribed species. We demonstrate that geometric morphometric techniques analysing wing shape represent a promising approach for discriminating between morphologically cryptic taxa of the B. dorsalis species complex.