Aphid flight-track analysis in three dimensions using video techniques

Abstract. A technique is described for the three-dimensional analysis of the flight paths of small insects using two video cameras placed alongside each other with the optical axes coincident at a point some distance beyond the area of interest. The video signals were mixed and a time base introduced before recording the superimposed images from both cameras on a single VCR. With suitable lighting and a black background, flying aphids appeared on the monitor as double, bright images on a dark background. The distance between the two images was inversely proportional to the distance of the aphid from the camera lenses. A calibration grid was used to insert the correct parameters into software designed to provide the x (vertical), y (horizontal) and z (distance from the cameras) Cartesian co-ordinates for a flying insect and to calculate the distances flown, flight speed and turning parameters. The advantages of the system are that it is designed for a single VCR and monitor, provides automatic synchrony between camera signals and can examine a larger visual arena than screen-splitting methods. It operates with insects as small as aphids, and wind-tunnel studies on the black bean aphid, Aphis fabae Scop., showed that some flight parameters (for the final one-second approach to a visually attractive landing platform) differed according to whether wind was present or not. Thus, ground speed and distance moved differed significantly but turning parameters were unchanged. In addition, flight trajectory on the approach to landing depended upon initial direction of flight and the presence of wind.     

Honeybees can discriminate between Monet and Picasso paintings

Honeybees (Apis mellifera) have remarkable visual learning and discrimination abilities that extend beyond learning simple colours, shapes or patterns. They can discriminate landscape scenes, types of flowers, and even human faces. This suggests that in spite of their small brain, honeybees have a highly developed capacity for processing complex visual information, comparable in many respects to vertebrates. Here, we investigated whether this capacity extends to complex images that humans distinguish on the basis of artistic style: Impressionist paintings by Monet and Cubist paintings by Picasso. We show that honeybees learned to simultaneously discriminate between five different Monet and Picasso paintings, and that they do not rely on luminance, colour, or spatial frequency information for discrimination. When presented with novel paintings of the same style, the bees even demonstrated some ability to generalize. This suggests that honeybees are able to discriminate Monet paintings from Picasso ones by extracting and learning the characteristic visual information inherent in each painting style. Our study further suggests that discrimination of artistic styles is not a higher cognitive function that is unique to humans, but simply due to the capacity of animals—from insects to humans—to extract and categorize the visual characteristics of complex images.     

A counting method for the number of Sternolophus rufipes and Hydrochara affinis in a noisy trap image

A quantitative survey of water scavenger beetles Sternolophus rufipes and Hydrochara affinis in paddy fields is essential not only for evaluating the impact of climate change on ecosystems but also for quantifying the stability of paddy fields. Many researchers classify insects in insect traps visually and manually count the number of individuals in each species. This manual survey method is time-consuming, fatiguing, and tedious. In this paper, we present a simple method to classify and count beetles in noisy trap images. The proposed method uses the beetles' body size and spots made by the light reflecting off the backs of beetles. We verify the method using images of beetles attached to the insect trap. The results demonstrate that the number of individuals in each species as counted by the proposed method and the manually counted number are statistically identical, which means that our method is sufficient to replace the existing manual counting method. Additionally, we briefly discuss the limitations of this counting method and ideas that could complement them.     

Sequence-independent characterization of viruses based on the pattern of viral small RNAs produced by the host

Virus surveillance in vector insects is potentially of great benefit to public health. Large-scale sequencing of small and long RNAs has previously been used to detect viruses, but without any formal comparison of different strategies. Furthermore, the identification of viral sequences largely depends on similarity searches against reference databases. Here, we developed a sequence-independent strategy based on virus-derived small RNAs produced by the host response, such as the RNA interference pathway. In insects, we compared sequences of small and long RNAs, demonstrating that viral sequences are enriched in the small RNA fraction. We also noted that the small RNA size profile is a unique signature for each virus and can be used to identify novel viral sequences without known relatives in reference databases. Using this strategy, we characterized six novel viruses in the viromes of laboratory fruit flies and wild populations of two insect vectors: mosquitoes and sandflies. We also show that the small RNA profile could be used to infer viral tropism for ovaries among other aspects of virus biology. Additionally, our results suggest that virus detection utilizing small RNAs can also be applied to vertebrates, although not as efficiently as to plants and insects.     

InsectCV: A system for insect detection in the lab from trap images

Advances in artificial intelligence, computer vision, and high-performance computing have enabled the creation of efficient solutions to monitor pests and identify plant diseases. In this context, we present InsectCV, a system for automatic insect detection in the lab from scanned trap images. This study considered the use of Moericke-type traps to capture insects in outdoor environments. Each sample can contain hundreds of insects of interest, such as aphids, parasitoids, thrips, and flies. The presence of debris, superimposed objects, and insects in varied poses is also common. To develop this solution, we used a set of 209 grayscale images containing 17,908 labeled insects. We applied the Mask R-CNN method to generate the model and created three web services for the image inference. The model training contemplated transfer learning and data augmentation techniques. This approach defined two new parameters to adjust the ratio of false positive by class, and change the lengths of the anchor side of the Region Proposal Network, improving the accuracy in the detection of small objects. The model validation used a total of 580 images obtained from field exposed traps located at Coxilha, and Passo Fundo, north of Rio Grande do Sul State, during wheat crop season in 2019 and 2020. Compared to manual counting, the coefficients of determination (R² = 0.81 for aphids and R² = 0.78 for parasitoids) show a good-fitting model to identify the fluctuation of population levels for these insects, presenting tiny deviations of the growth curve in the initial phases, and in the maintenance of the curve shape. In samples with hundreds of insects and debris that generate more connections or overlaps, model performance was affected due to the increase in false negatives. Comparative tests between InsectCV and manual counting performed by a specialist suggest that the system is sufficiently accurate to guide warning systems for integrated pest management of aphids. We also discussed the implications of adopting this tool and the gaps that require further development.     

Visualization of growth signal transduction cascades in living cells with genetically encoded probes based on Förster resonance energy transfer

Fluorescence probes based on the principle of Förster resonance energy transfer (FRET) have shed new light on our understanding of signal transduction cascades. Among them, unimolecular FRET probes containing fluorescence proteins are rapidly increasing in number because these genetically encoded probes can be easily loaded into living cells and allow simple acquisition of FRET images. We have developed probes for small GTPases, tyrosine kinases, serine–threonine kinases and phosphoinositides. Images obtained with these probes have revealed that membrane protrusions such as nascent lamellipodia or neurites provide an active signalling platform in the growth factor-stimulated cells.     

Use of the FlowCAM for semi-automated recognition and enumeration of red tide cells (Karenia brevis) in natural plankton samples

Early detection is the most effective way to mitigate the effects of harmful algal blooms (HAB). Cell counts based on examination of microplankton samples using settling chambers and visual inspection with an inverted microscope are tedious and time consuming, and counting precision is generally poor at low cell densities. The FlowCAM is a continuous imaging flow cytometer designed to characterize particles in the microplankton size range (20–200 μm diameter). In this study we examined the ability of the FlowCAM to improve routine monitoring protocols for HAB species by automatically recording information on size and fluorescence per cell. This will eliminate the need to examine cells outside the ranges of these measurements for our target species, Karenia brevis. We also tested the ability of image comparison software to match images of cells in mixed assemblages to images of the target species. For simple mixtures of cultured dinoflagellates, the ability of the image matching software to discriminate target cells varied greatly depending on how similar the two species were in size and shape. When target cells were added to natural plankton samples, the image recognition software correctly identified 80–90% of the target cells, but misidentified 20–50% of non-target cells in the size range of the target species. We conclude that the FlowCAM is less tedious and time-consuming than microscopy, allowing for examination of more cells for greater counting precision. The cell recognition software helps reduce the numbers of cells that must be screened, but images must still be examined by a trained operator to identify the HAB species of interest.     

Contribution of small insects to pollination of common buckwheat, a distylous crop

Crop pollination by animals is an essential ecosystem service. Among animal-pollinated crops, distylous plants strongly depend on animal pollination. In distylous pollination systems, pollinator species are usually limited, although flowers of some distylous plants are visited by diverse animals. We studied the pollination biology of common buckwheat ( Fagopyrum esculentum ), a distylous crop mainly pollinated by honeybees and visited by many insect species, to evaluate the effects of non-honeybee species on pollination services. We focused on insects smaller than honeybees to determine their contribution to pollination. We applied pollination treatments with bags of coarse mesh to exclude flower visits by honeybees and larger insects and compared the seed set of bagged plants with that of untreated plants for pin and thrum flower morphs. We found a great reduction of seed set only in bagged pin flowers. We also confirmed that small insects, including ants, bees, wasps and flies, carried pin-morph pollen. These small insects transfer pollen from the short anthers of pin flowers to the short styles of thrum flowers, leading to sufficient seed set in thrum flowers. Consequently, small, non-honeybee insects have the potential to maintain at least half of the yield of this honeybee-dependent distylous crop.     

Rapid photogrammetry of morphological traits of free-ranging moths

1. Photogrammetric studies of free-ranging animals are limited to mammals and birds. Recent advances in insect photogrammetry, including 3D imaging, are entirely associated with museum specimens.
2. We present a rapid, simple, accurate, and inexpensive morphometric method targeting thousands of free-ranging insects attracted to light screens using images taken without collecting a specimen or even constraining the individual in any manner. A reference grid printed on the screen is used to calibrate the images for shape and size without prior knowledge of the camera-subject configuration. The method requires only inexpensive, off-the-shelf, consumer equipment, and freely available programming (R statistical language) and image processing (ImageMagick) tools.
3. We demonstrate the efficacy of the method using a dataset of 3675 images of free-ranging hawkmoths (Lepidoptera:Sphingidae) imaged in natural repose on a screen. We show that this method introduces no bias and has a high degree of correspondence with traditional morphometry using collected specimens. We also propose error metrics, which quantify the calibration quality and identify images with poor data.
4. Although this method is particularly suited for the hyperdiverse moth community, which dominates the dynamics of many terrestrial ecosystems, it can be used for other phototropic taxa identifiable on an image to (morpho)-species. It will help in accumulating reliable trait data from hundreds of thousands of individual insects without any expenditure on specimen collection. It is particularly suited for studies which require multi-epoch, multi-locate sampling like investigations into ecosystem stability, climate change, and community assembly.

     

Mixtures of conditional gaussian scale mixtures applied to multiscale image representations

We present a probabilistic model for natural images that is based on mixtures of Gaussian scale mixtures and a simple multiscale representation. We show that it is able to generate images with interesting higher-order correlations when trained on natural images or samples from an occlusion-based model. More importantly, our multiscale model allows for a principled evaluation. While it is easy to generate visually appealing images, we demonstrate that our model also yields the best performance reported to date when evaluated with respect to the cross-entropy rate, a measure tightly linked to the average log-likelihood. The ability to quantitatively evaluate our model differentiates it from other multiscale models, for which evaluation of these kinds of measures is usually intractable.     

Benefits conferred by “timid” ants: active anti-herbivore protection of the rainforest tree Leonardoxa africana by the minute ant Petalomyrmex phylax

In this study, we demonstrate that an important benefit provided by the small host-specific ant Petalomyrmex phylax to its host plant Leonardoxa africana is efficient protection against herbivores. We estimate that in the absence of ants, insect herbivory would reduce the leaf area by about one-third. This contributes considerably to the fitness of the plant. Our estimates take into account not only direct damage, such as removal of leaf surface by chewing insects, but also the effects of sucking insects on leaf growth and expansion. Sucking insects are numerically predominant in this system, and the hitherto cryptic effects of ant protection against the growth-reducing effects of sucking insects accounted for half of the total estimated benefit of ant protection. We propose that the small size of workers confers a distinct advantage in this system. Assuming that resource limitation implies a trade off between size and number of ants, and given the small size of phytophagous insects that attack Leonardoxa, we conclude that fine-grained patrolling by a large number of small workers maximises protection of young leaves of this plant. Since herbivores are small and must complete their development on the young leaves of Leonardoxa, and since a high patrolling density is required for a fine-grained search for these enemies, numerous small ants should provide the most effective protection of young leaves of Leonardoxa. We also discuss other factors that may have influenced worker size in this ant. Received: 1 September 1996 / Accepted: 2 June 1997     

Probing f-actin flow by tracking shape fluctuations of radial bundles in lamellipodia of motile cells

We examined the dynamics of radial actin bundles based on time-lapse movies of polarized light images of living neuronal growth cones. Using a highly sensitive computer vision algorithm for tracking, we analyzed the small shape fluctuations of radial actin bundles that otherwise remained stationary in their positions in the growth cone lamellipodium. Using the tracking software, we selected target points on radial bundles and measured both the local bundle orientations and the lateral displacements between consecutive movie frames. We found that the local orientation and the lateral displacement of a target point are correlated. The correlation can be explained using a simple geometric relationship between the lateral travel of tilted actin bundles and the retrograde flow of f-actin structures. Once this relationship has been established, we have turned the table and used the radial bundles as probes to measure the velocity field of f-actin flow. We have generated a detailed map of the complex retrograde flow pattern throughout the lamellipodium. Such two-dimensional flow maps will give new insights into the mechanisms responsible for f-actin-mediated cell motility and growth.     

Validity of thermal ramping assays used to assess thermal tolerance in arthropods

Proper assessment of environmental resistance of animals is critical for the ability of researchers to understand how variation in environmental conditions influence population and species abundance. This is also the case for studies of upper thermal limits in insects, where researchers studying animals under laboratory conditions must select appropriate methodology on which conclusions can be drawn. Ideally these methods should precisely estimate the trait of interest and also be biological meaningful. In an attempt to develop such tests it has been proposed that thermal ramping assays are useful assays for small insects because they incorporate an ecologically relevant gradual temperature change. However, recent model-based papers have suggested that estimates of thermal resistance may be strongly confounded by simultaneous starvation and dehydration stress. In the present study we empirically test these model predictions using two sets of independent experiments. We clearly demonstrate that results from ramping assays of small insects (Drosophila melanogaster) are not compromised by starvation- or dehydration-stress. Firstly we show that the mild disturbance of water and energy balance of D. melanogaster experienced during the ramping tests does not confound heat tolerance estimates. Secondly we show that flies pre-exposed to starvation and dehydration have "normal" heat tolerance and that resistance to heat stress is independent of the energetic and water status of the flies. On the basis of our results we discuss the assumptions used in recent model papers and present arguments as to why the ramping assay is both a valid and ecologically relevant way to measure thermal resistance in insects.     

The influence of freshwater-lake subsidies on invertebrates occupying terrestrial vegetation

Studies investigating effects of aquatic-derived resource subsidies have often found large effects on terrestrial systems. Those studies have mostly been performed on effects of subsidies derived from oceanic and riverine systems, and very few have considered effects of subsidies from freshwater lakes. However, since lakes can produce large quantities of emergent aquatic insects that end up on nearby land, it is likely that also freshwater-lake subsidies influence terrestrial systems. We performed sweep-net collections of aquatic and terrestrial invertebrates at varying distances from the shore on vegetation of islands of varying size, in two freshwater lakes in northern Sweden, as well as on the surrounding mainland. We found that the amounts of aquatic insects on terrestrial vegetation decreased with distance from the shore, and that they were the most abundant on small islands, presumably because small islands have a higher perimeter-to-area ratio. Web-building spiders responded positively to the aquatic subsidy by being the most abundant on small islands and by showing a positive relationship with aquatic insect biomass. However, distance from the shore showed no effects on the spiders. Our results strongly support the view that terrestrial systems are subsidized by lakes, and indicate that freshwater-lake subsidies are important for terrestrial invertebrate community structure on adjacent land. Further, our study shows that ecosystems should be treated as interdependent, not as self-contained units, and may as such be important for an increased understanding of the nature and importance of resource flows across ecosystem boundaries.     

Visual Recognition Software for Binary Classification and Its Application to Spruce Pollen Identification

Discriminating between black and white spruce (Picea mariana and Picea glauca) is a difficult palynological classification problem that, if solved, would provide valuable data for paleoclimate reconstructions. We developed an open-source visual recognition software (ARLO, Automated Recognition with Layered Optimization) capable of differentiating between these two species at an accuracy on par with human experts. The system applies pattern recognition and machine learning to the analysis of pollen images and discovers general-purpose image features, defined by simple features of lines and grids of pixels taken at different dimensions, size, spacing, and resolution. It adapts to a given problem by searching for the most effective combination of both feature representation and learning strategy. This results in a powerful and flexible framework for image classification. We worked with images acquired using an automated slide scanner. We first applied a hash-based “pollen spotting” model to segment pollen grains from the slide background. We next tested ARLO’s ability to reconstruct black to white spruce pollen ratios using artificially constructed slides of known ratios. We then developed a more scalable hash-based method of image analysis that was able to distinguish between the pollen of black and white spruce with an estimated accuracy of 83.61%, comparable to human expert performance. Our results demonstrate the capability of machine learning systems to automate challenging taxonomic classifications in pollen analysis, and our success with simple image representations suggests that our approach is generalizable to many other object recognition problems.     

Motion vectors and deep neural networks for video camera traps

Commercial camera traps are usually triggered by a Passive Infra-Red (PIR) motion sensor necessitating a delay between triggering and the image being captured. This often seriously limits the ability to record images of small and fast moving animals. It also results in many “empty” images, e.g., owing to moving foliage against a background of different temperature. In this paper we detail a new triggering mechanism based solely on the camera sensor. This is intended for use by citizen scientists and for deployment on an affordable, compact, low-power Raspberry Pi computer (RPi). Our system introduces a video frame filtering pipeline consisting of movement and image-based processing. This makes use of Machine Learning (ML) feasible on a live camera stream on an RPi. We describe our free and open-source software implementation of the system; introduce a suitable ecology efficiency measure that mediates between specificity and recall; provide ground-truth for a video clip collection from camera traps; and evaluate the effectiveness of our system thoroughly. Overall, our video camera trap turns out to be robust and effective.     

The evolutionary ecology of detritus feeding in the larvae of freshwater Diptera

Detritus (dead organic matter), largely of terrestrial origin, is superabundant in inland waters but because of its indigestible nature, would appear to be a poor food source for animals. Yet this unpromising material is widely used as food and indeed can be viewed as a defining characteristic of the freshwater environment. We here explore the relationships among animals, detritus and its associated micro-organism decomposers, taking a functional approach. We pose questions about interrelationships and attempt to arrive at new insights by disentangling them from an adaptive point of view. To do this we have been careful in selecting the habitats for detailed consideration. Rain pools on rock surfaces in tropical Africa and pools on peat moorland in the UK were chosen. Both examples have a relatively simple community structure and hence offer the prospect of achieving our aim. As model organisms for study we focus principally on the aquatic stages of selected holometabolous insects; that is, selected genera of the universally common midges, Ceratopogonidae and Chironomidae. We approach these case studies from an evolutionary ecology perspective and see detritus as a simple template upon which a beautiful complex of adaptations can evolve.     

Comparison of image analysis software packages in the assessment of adhesion of microorganisms to mucosal epithelium using confocal laser scanning microscopy

We have compared current image analysis software packages in order to find the most useful one for assessing microbial adhesion and inhibition of adhesion to tissue sections. We have used organisms of different sizes, the bacterium Helicobacter pylori and the yeast Candida albicans. Adhesion of FITC-labelled H. pylori and C. albicans was assessed by confocal microscopy. Four different Image analysis software packages, NIH-Image, IP Lab, Image Pro+, and Metamorph, were compared for their ability to quantify adhesion of the two organisms and several quantification methods were devised for each package. For both organisms, the dynamic range that could be detected by the software packages was 1x10(6)-1x10(9) cells/ml. Of the four software packages tested, our results showed that Metamorph software, using our 'Region of Interest' method, with the software's 'Standard Area Method' of counting, was the most suitable for quantifying adhesion of both organisms because of its unique ability to separate clumps of microbial cells. Moreover, fewer steps were required. By pre-incubating H. pylori with the glycoconjugate Lewis b-HSA, an inhibition of binding of 48.8% was achieved using 250 mug/ml Lewis b-HSA. The method we have devised using Metamorph software, provides a simple, quick and accurate way of quantifying adhesion and inhibition of adhesion of microbial cells to the epithelial surface of tissue sections. The method can be applied to organisms ranging in size from small bacteria to larger yeast cells.     

Software for minimalistic data management in large camera trap studies

The use of camera traps is now widespread and their importance in wildlife studies is well understood. Camera trap studies can produce millions of photographs and there is a need for a software to help manage photographs efficiently. In this paper, we describe a software system that was built to successfully manage a large behavioral camera trap study that produced more than a million photographs. We describe the software architecture and the design decisions that shaped the evolution of the program over the study's three year period. The software system has the ability to automatically extract metadata from images, and add customized metadata to the images in a standardized format. The software system can be installed as a standalone application on popular operating systems. It is minimalistic, scalable and extendable so that it can be used by small teams or individual researchers for a broad variety of camera trap studies.