Automated processing of webcam images for phenological classification

Abstract

Along with the global climate change, there is an increasing interest for its effect on phenological patterns such as start and end of the growing season. Scientific digital webcams are used for this purpose taking every day one or more images from the same natural motive showing for example trees or grassland sites. To derive phenological patterns from the webcam images, regions of interest are manually defined on these images by an expert and subsequently a time series of percentage greenness is derived and analyzed with respect to structural changes. While this standard approach leads to satisfying results and allows to determine dates of phenological change points, it is associated with a considerable amount of manual work and is therefore constrained to a limited number of webcams only. In particular, this forbids to apply the phenological analysis to a large network of publicly accessible webcams in order to capture spatial phenological variation. In order to be able to scale up the analysis to several hundreds or thousands of webcams, we propose and evaluate two automated alternatives for the definition of regions of interest, allowing for efficient analyses of webcam images. A semi-supervised approach selects pixels based on the correlation of the pixels’ time series of percentage greenness with a few prototype pixels. An unsupervised approach clusters pixels based on scores of a singular value decomposition. We show for a scientific webcam that the resulting regions of interest are at least as informative as those chosen by an expert with the advantage that no manual action is required. Additionally, we show that the methods can even be applied to publicly available webcams accessed via the internet yielding interesting partitions of the analyzed images. Finally, we show that the methods are suitable for the intended big data applications by analyzing 13988 webcams from the AMOS database. All developed methods are implemented in the statistical software package R and publicly available in the R package . Executable example code is provided as supplementary material.

Remote sensing of temperate and boreal forest phenology: A review of progress, challenges and opportunities in the intercomparison of in-situ and satellite phenological metrics

“…As an example, the Phenocam network (Richardson et al, 2007) monitors hundreds of sites (mainly in North America); but there are a few more networks fully operational in Europe, Asia and Australia (Moore et al, 2016) and Brazil (Alberton et al, 2017;Alberton et al, 2019). Besides this, studies have being investigating the potential of using traffic and security cameras as phenocams (Morris et al, 2013;Bothmann et al, 2017), which could increase dramatically the number of sampling locations and diminish installation costs. Furthermore, additional instrumentation is usually available (e.g., meteorological stations), which can supply with extra ground-truth data (Moore et al, 2016), which in turn could: 1) feed models such as the Quadratic, which requires the calculation of the accumulated growing degree-days (White et al, 2009); and 2) enable comparisons with meteorological data (Bradley et al, 2010).…”

A Proposal for a Forest Digital Twin Framework and Its Perspectives

“…(1) the risk of reversals impacting the asset value provided by an ecosystem service shared with a potential buyer (e.g., risk of wind storm hindering the forest's potential of carbon sequestration); (2) signals of the forest's health degradation for the scientific community (e.g., increasing episodes of hydraulic failure, carbon starvation, insects, and pathogens). Forest management Field measurements and metadata [53] Hydrological basin parameters Remote sensing, IoT [54] Species diversity Remote sensing [55,56] Stand structure Remote sensing, IoT [57][58][59][60][61][62] Weather IoT [63] Wildlife and herbivores IoT [64,65] Ideally, by adopting common standards for tracking and reporting, multiple implementations of FDT could build a global network like Fluxnet [49], thus, generating a distinctive globally data-driven repository for the scientific community. Furthermore, reliable data on threats affecting forest ecosystems can improve risk awareness and foster the implementation of mitigation actions [66].…”

Continuous Monitoring of Tree Responses to Climate Change for Smart Forestry: A Cybernetic Web of Trees

“…A number of traits are good indicators of tree responses to resource availability, or biotic disturbance, and data processed by software platforms can be readily converted into descriptions of these traits. Integrating image processing (e.g., scientific digital webcams; Bothmann et al 2017) with functional monitoring (e.g., sap flow gauges; Flo et al 2019) provides an example of how different sensors can be linked to address rapid dynamics in plant response to environmental changes. The fast development of advanced equipment and the vast amount of generated data may allow innovative data-driven approaches to replace traditional hypothesis-driven analyses, providing new insights on forest ecophysiology by means of artificial intelligence, e.g., machine learning approaches (Torresan et al 2021).…”