Dataset Information

Automatic classification of fine-scale mountain vegetation based on mountain altitudinal belt.

ABSTRACT: Vegetation mapping is of considerable significance to both geoscience and mountain ecology, and the improved resolution of remote sensing images makes it possible to map vegetation at a finer scale. While the automatic classification of vegetation has gradually become a research hotspot, real-time and rapid collection of samples has become a bottleneck. How to achieve fine-scale classification and automatic sample selection at the same time needs further study. Stratified sampling based on appropriate prior knowledge is an effective sampling method for geospatial objects. Therefore, based on the idea of stratified sampling, this paper used the following three steps to realize the automatic selection of representative samples and classification of fine-scale mountain vegetation: 1) using Mountain Altitudinal Belt (MAB) distribution information to stratify the study area into multiple vegetation belts; 2) selecting and correcting samples through iterative clustering at each belt automatically; 3) using RF (Random Forest) classifier with strong robustness to achieve automatic classification. The average sample accuracy of nine vegetation formations was 0.933, and the total accuracy of the classification result was 92.2%, with the kappa coefficient of 0.910. The results showed that this method could automatically select high-quality samples and obtain a high-accuracy vegetation map. Compared with the traditional vegetation mapping method, this method greatly improved the efficiency, which is of great significance for the fine-scale mountain vegetation mapping in large-scale areas.

SUBMITTER: Zhang J

PROVIDER: S-EPMC7447069 | biostudies-literature | 2020

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Automatic classification of fine-scale mountain vegetation based on mountain altitudinal belt.

Zhang Junyao J Yao Yonghui Y Suo Nandongzhu N

PloS one 20200825 8

Vegetation mapping is of considerable significance to both geoscience and mountain ecology, and the improved resolution of remote sensing images makes it possible to map vegetation at a finer scale. While the automatic classification of vegetation has gradually become a research hotspot, real-time and rapid collection of samples has become a bottleneck. How to achieve fine-scale classification and automatic sample selection at the same time needs further study. Stratified sampling based on appro ...[more]

PMID: 32841269

Similar Datasets

Project description:BACKGROUND:Molecular studies in social mammals rarely compare the inferences gained from genetic analyses with field information, especially in the context of dispersal. In this study, we used genetic data to elucidate sex-specific dispersal dynamics in the Virunga Massif mountain gorilla population (Gorilla beringei beringei), a primate species characterized by routine male and female dispersal from stable mixed-sex social groups. Specifically, we conducted spatial genetic structure analyses for each sex and linked our genetically-based observations with some key demographic and behavioural data from this population. RESULTS:To investigate the spatial genetic structure of mountain gorillas, we analysed the genotypes of 193 mature individuals at 11 microsatellite loci by means of isolation-by-distance and spatial autocorrelation analyses. Although not all males and females disperse, female gorillas displayed an isolation-by-distance pattern among groups and a signal of dispersal at short distances from their natal group based on spatial autocorrelation analyses. In contrast, male genotypes were not correlated with spatial distance, thus suggesting a larger mean dispersal distance for males as compared to females. Both within sex and mixed-sex pairs were on average genetically more related within groups than among groups. CONCLUSIONS:Our study provides evidence for an intersexual difference in dispersal distance in the mountain gorilla. Overall, it stresses the importance of investigating spatial genetic structure patterns on a sex-specific basis to better understand the dispersal dynamics of the species under investigation. It is currently poorly understood why some male and female gorillas disperse while others remain in the natal group. Our results on average relatedness within and across groups confirm that groups often contain close relatives. While inbreeding avoidance may play a role in driving female dispersal, we note that more detailed dyadic genetic analyses are needed to shed light on the role of inbreeding avoidance as an ultimate cause of female dispersal in mountain gorillas.

Project description:Understanding how land-use changes affect different facets of plant biodiversity in seminatural European grasslands is of particular importance for biodiversity conservation. As conclusions of previous experimental or synchronic observational studies did not converge toward a general agreement, assessing the recent trends in vegetation change in various grassland systems using a diachronic approach is needed. In this resurvey study, we investigated the recent changes in grassland vegetation of the French Jura Mountains, a region with a long tradition of pastoralism. We compared the floristic composition of 150 grassland plots recorded between 1990 and 2000 with new relevés made in 2012 on the same plots. We considered taxonomic, phylogenetic and functional diversity as well as ecological characteristics of the plant communities derived from ecological indicator values and life strategies of the species. PCA of the floristic composition revealed a significant general trend linked to the sampling year. Wilcoxon paired tests showed that contemporary communities were generally more dominated by grass species and presented a higher tolerance to defoliation, a higher pastoral value, and a higher nutrient indicator value. Comparisons revealed a decrease in phylogenetic and functional diversity. By contrast, local species richness has slightly increased. The intensity of change in species composition, measured by Hellinger distance between pairs of relevés, was dependent on neither the time lag between the two surveys, the author of the first relevé nor its location or elevation. The most important changes were observed in grasslands that previously presented low pastoral value, low grass cover, low tolerance to defoliation, and high proportion of stress-tolerant species. This trend was likely linked to the intensification of grassland management reported in the region, with a parallel increase in mowing frequency, grazing pressure, and fertilization level. More restrictive specifications should be applied to agricultural practices to avoid overexploitation of mountain species-rich grasslands and its negative consequences on their biodiversity and resilience.

Project description:Automatic species classification of birds from their sound is a computational tool of increasing importance in ecology, conservation monitoring and vocal communication studies. To make classification useful in practice, it is crucial to improve its accuracy while ensuring that it can run at big data scales. Many approaches use acoustic measures based on spectrogram-type data, such as the Mel-frequency cepstral coefficient (MFCC) features which represent a manually-designed summary of spectral information. However, recent work in machine learning has demonstrated that features learnt automatically from data can often outperform manually-designed feature transforms. Feature learning can be performed at large scale and "unsupervised", meaning it requires no manual data labelling, yet it can improve performance on "supervised" tasks such as classification. In this work we introduce a technique for feature learning from large volumes of bird sound recordings, inspired by techniques that have proven useful in other domains. We experimentally compare twelve different feature representations derived from the Mel spectrum (of which six use this technique), using four large and diverse databases of bird vocalisations, classified using a random forest classifier. We demonstrate that in our classification tasks, MFCCs can often lead to worse performance than the raw Mel spectral data from which they are derived. Conversely, we demonstrate that unsupervised feature learning provides a substantial boost over MFCCs and Mel spectra without adding computational complexity after the model has been trained. The boost is particularly notable for single-label classification tasks at large scale. The spectro-temporal activations learned through our procedure resemble spectro-temporal receptive fields calculated from avian primary auditory forebrain. However, for one of our datasets, which contains substantial audio data but few annotations, increased performance is not discernible. We study the interaction between dataset characteristics and choice of feature representation through further empirical analysis.

Project description:BackgroundUnderstanding altitudinal patterns of biological diversity and their underlying mechanisms is critically important for biodiversity conservation in mountainous regions. The contribution of area to plant diversity patterns is widely acknowledged and may mask the effects of other determinant factors. In this context, it is important to examine altitudinal patterns of corrected taxon richness by eliminating the area effect. Here we adopt two methods to correct observed taxon richness: a power-law relationship between richness and area, hereafter "method 1"; and richness counted in equal-area altitudinal bands, hereafter "method 2". We compare these two methods on the Jade Dragon Snow Mountain, which is the nearest large-scale altitudinal gradient to the Equator in the Northern Hemisphere.ResultsWe find that seed plant species richness, genus richness, family richness, and species richness of trees, shrubs, herbs and Groups I-III (species with elevational range size <150, between 150 and 500, and >500 m, respectively) display distinct hump-shaped patterns along the equal-elevation altitudinal gradient. The corrected taxon richness based on method 2 (TRcor2) also shows hump-shaped patterns for all plant groups, while the one based on method 1 (TRcor1) does not. As for the abiotic factors influencing the patterns, mean annual temperature, mean annual precipitation, and mid-domain effect explain a larger part of the variation in TRcor2 than in TRcor1.ConclusionsIn conclusion, for biodiversity patterns on the Jade Dragon Snow Mountain, method 2 preserves the significant influences of abiotic factors to the greatest degree while eliminating the area effect. Our results thus reveal that although the classical method 1 has earned more attention and approval in previous research, method 2 can perform better under certain circumstances. We not only confirm the essential contribution of method 1 in community ecology, but also highlight the significant role of method 2 in eliminating the area effect, and call for more application of method 2 in further macroecological studies.

Dataset Information

Automatic classification of fine-scale mountain vegetation based on mountain altitudinal belt.

Publications

Automatic classification of fine-scale mountain vegetation based on mountain altitudinal belt.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets