Object-based Analysis Research Articles

Object-based analysis of simulated thunderstorms in Switzerland: application and validation of automated thunderstorm tracking with simulation data

Abstract. We present a feasibility study for an object-based method to characterise thunderstorm properties in simulation data from convection-permitting weather models. An existing thunderstorm tracker, the Thunderstorm Identification, Tracking, Analysis and Nowcasting (TITAN) algorithm, was applied to thunderstorms simulated by the Advanced Research Weather Research and Forecasting (AR-WRF) weather model at convection-permitting resolution for a domain centred on Switzerland. Three WRF microphysics parameterisations were tested. The results are compared to independent radar-based observations of thunderstorms derived using the MeteoSwiss Thunderstorms Radar Tracking (TRT) algorithm. TRT was specifically designed to track thunderstorms over the complex Alpine topography of Switzerland. The object-based approach produces statistics on the simulated thunderstorms that can be compared to object-based observation data. The results indicate that the simulations underestimated the occurrence of severe and very large hail compared to the observations. Other properties, including the number of storm cells per day, geographical storm hotspots, thunderstorm diurnal cycles, and storm movement directions and velocities, provide a reasonable match to the observations, which shows the feasibility of the technique for characterisation of simulated thunderstorms over complex terrain.

Machine Learning Classification and Accuracy Assessment from High-Resolution Images of Coastal Wetlands

High-resolution images obtained by multispectral cameras mounted on Unmanned Aerial Vehicles (UAVs) are helping to capture the heterogeneity of the environment in images that can be discretized in categories during a classification process. Currently, there is an increasing use of supervised machine learning (ML) classifiers to retrieve accurate results using scarce datasets with samples with non-linear relationships. We compared the accuracies of two ML classifiers using a pixel and object analysis approach in six coastal wetland sites. The results show that the Random Forest (RF) performs better than K-Nearest Neighbors (KNN) algorithm in the classification of pixels and objects and the classification based on pixel analysis is slightly better than the object-based analysis. The agreement between the classifications of objects and pixels is higher in Random Forest. This is likely due to the heterogeneity of the study areas, where pixel-based classifications are most appropriate. In addition, from an ecological perspective, as these wetlands are heterogeneous, the pixel-based classification reflects a more realistic interpretation of plant community distribution.

Evaluating the impact of model resolutions and cumulus parameterization on precipitation in NU-WRF: A case study in the Central Great Plains

Regional climate models are expected to exhibit improved skill at finer spatial resolutions due to improved representation of land surface heterogeneity. However, at spatial scales between 1 and 10 km (grey scales), these improvements are often illusive due to the competing benefits from spatial resolution and cumulus parameterization. This study provides insights into the impact of model resolution and cumulus parameterization on precipitation prediction in the Central Great Plains by using an object-based evaluation method. Our results show limited improvement solely from finer resolution but larger improvement without using the cumulus scheme at a 4-km resolution. Compared to traditional evaluation methods, the object-based analysis shows that without the cumulus scheme the spatial properties of precipitation are better represented. In contrast, all model configurations show a dry bias in precipitation days and a tendency to produce widespread precipitation but with fewer hours with precipitation which indicates other shortcomings in the model.

Characterizing Small-Town Development Using Very High Resolution Imagery within Remote Rural Settings of Mozambique

While remotely sensed images of various resolutions have been widely used in identifying changes in urban and peri-urban environments, only very high resolution (VHR) imagery is capable of providing the information needed for understanding the changes taking place in remote rural environments, due to the small footprints and low density of man-made structures in these settings. However, limited by data availability, mapping man-made structures and conducting subsequent change detections in remote areas are typically challenging and thus require a certain level of flexibility in algorithm design that takes into account the specific environmental and image conditions. In this study, we mapped all buildings and corrals for two remote villages in Mozambique based on two single-date VHR images that were taken in 2004 and 2012, respectively. Our algorithm takes advantage of the presence of shadows and, through a fusion of both spectra- and object-based analysis techniques, is able to differentiate buildings with metal and thatch roofs with high accuracy (overall accuracy of 86% and 94% for 2004 and 2012, respectively). The comparison of the mapping results between 2004 and 2012 reveals multiple lines of evidence suggesting that both villages, while differing in many aspects, have experienced substantial increases in the economic status. As a case study, our project demonstrates the capability of a coupling of VHR imagery with locally adjusted classification algorithms to infer the economic development of small, remote rural settlements.

Erratum to “Semantic Unsupervised Change Detection of Natural Land Cover With Multitemporal Object-Based Analysis on SAR Images” [Jul 21 5494-5514

In the above article <xref ref-type="bibr" rid="ref1" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[1]</xref> , the author affiliations were incorrectly listed. The correct affiliations are as follows:

Mapping tillage direction and contour farming by object-based analysis of UAV images

Tillage is a primary agricultural task that causes progressive soil movement and, consequently, severe erosion in sloping farmland, with a high impact on crop productivity, soil quality and landscape features. Governments and public administrations acknowledge the need to control tillage operations for a better soil conservation in vulnerable areas, recommending the practice of contour farming. Despite this importance, there is currently no mechanism for the effective and large-scale monitoring of tillage patterns at the field or regional scale. Accordingly, this research combined aerial images taken with unmanned aerial vehicles and object-based image analysis (OBIA) to develop an innovative OBIA4tillage procedure with three main objectives: (i) analysing plowed agricultural fields, identifying and mapping the tillage marks, and automatically computing the main direction of the tillage furrows, (ii) validating the procedure quality in different scenarios by evaluating the accuracy of the results as affected by the sensor used (visible-light vs. multispectral), background soil hue, and ground vegetation density; and (iii) mapping contour farming and non-contour farming areas as indicators of potential low and high soil erosion risk, respectively. Twenty olive parcels from two different locations with a wide range of tree sizes, soil hue, parcel shapes and land slopes were selected as model systems to develop and validate the procedure. The OBIA4tillage procedure produced tillage maps with very high accuracy for both RGB and multispectral images (R2 of 0.99 and 0.93, respectively), as obtained from the linear equation between estimated and ground-truth values. The results were similar in clear and dark soils (R2 of 0.96 in both cases), although there were notable differences between areas of dense ground vegetation or bare soil (R2 of 0.99 in both cases) and areas of medium vegetation cover (R2 of 0.81). The application of contour farming in the study region was moderate at location 1 (42.35% of the study area) but more widespread at location 2 (72.60% of the study area), which revealed the uneven involvement of the local farmers in the challenge of controlling soil erosion risks. The valuable information in the OBIA4tillage outputs can also be used to study tillage events at various spatial and temporal scales, feed soil erosion models and promote good agricultural and environmental conditions measures implemented at the parcel scale.

An Adaptive Tracking Algorithm for Convection in Simulated and Remote Sensing Data

AbstractA robust and computationally efficient object tracking algorithm is developed by incorporating various tracking techniques. Physical properties of the objects, such as brightness temperature or reflectivity, are not considered. Therefore, the algorithm is adaptable for tracking convection-like features in simulated data and remotely sensed two-dimensional images. In this algorithm, a first guess of the motion, estimated using the Fourier phase shift, is used to predict the candidates for matching. A disparity score is computed for each target–candidate pair. The disparity also incorporates overlapping criteria in the case of large objects. Then the Hungarian method is applied to identify the best pairs by minimizing the global disparity. The high-disparity pairs are unmatched, and their target and candidate are declared expired and newly initiated objects, respectively. They are tested for merger and split on the basis of their size and overlap with the other objects. The sensitivity of track duration is shown for different disparity and size thresholds. The paper highlights the algorithm’s ability to study convective life cycles using radar and simulated data over Darwin, Australia. The algorithm skillfully tracks individual convective cells (a few pixels in size) and large convective systems. The duration of tracks and cell size are found to be lognormally distributed over Darwin. The evolution of size and precipitation types of isolated convective cells is presented in the Lagrangian perspective. This algorithm is part of a vision for a modular platform [viz., TINT is not TITAN (TINT) and Tracking and Object-Based Analysis of Clouds (tobac)] that will evolve into a sustainable choice to analyze atmospheric features.

Atmospheric boundary layer height estimation from aerosol lidar: a new approach based on morphological image processing techniques

Abstract. The atmospheric boundary layer (ABL) represents the lowermost part of the atmosphere directly in contact with the Earth's surface. The estimation of its depth is of crucial importance in meteorology and for anthropogenic pollution studies. ABL height (ABLH) measurements are usually far from being adequate, both spatially and temporally. Thus, different remote sensing sources can be of great help in growing both the spatial and temporal ABLH measurement capabilities. To this aim, aerosol backscatter profiles are widely used as a proxy to retrieve the ABLH. Hence, the scientific community is making remarkable efforts in developing automatic ABLH retrieval algorithms applied to lidar observations. In this paper, we propose a ABLH estimation algorithm based on image processing techniques applied to the composite image of the total attenuated backscatter coefficient. A pre-processing step is applied to the composite total backscatter image based on morphological filters to properly set-up and adjust the image to detect edges. As final step, the detected edges are post-processed through both mathematical morphology and an object-based analysis. The performance of the proposed approach is assessed on real data acquired by two different lidar systems, deployed in Potenza (Italy) and Évora (Portugal), belonging to the European Aerosol Research Lidar Network (EARLINET). The proposed approach has shown higher performance than the benchmark consisting of some state-of-the-art ABLH estimation methods.

Measuring Nanoscale Distances by Structured Illumination Microscopy and Image Cross-Correlation Spectroscopy (SIM-ICCS).

Since the introduction of super-resolution microscopy, there has been growing interest in quantifying the nanoscale spatial distributions of fluorescent probes to better understand cellular processes and their interactions. One way to check if distributions are correlated or not is to perform colocalization analysis of multi-color acquisitions. Among all the possible methods available to study and quantify the colocalization between multicolor images, there is image cross-correlation spectroscopy (ICCS). The main advantage of ICCS, in comparison with other co-localization techniques, is that it does not require pre-segmentation of the sample into single objects. Here we show that the combination of structured illumination microscopy (SIM) with ICCS (SIM-ICCS) is a simple approach to quantify colocalization and measure nanoscale distances from multi-color SIM images. We validate the SIM-ICCS analysis on SIM images of optical nanorulers, DNA-origami-based model samples containing fluorophores of different colors at a distance of 80 nm. The SIM-ICCS analysis is compared with an object-based analysis performed on the same samples. Finally, we show that SIM-ICCS can be used to quantify the nanoscale spatial distribution of functional nuclear sites in fixed cells.

Remote sensing image segmentation advances: A meta-analysis

The advances in remote sensing sensors during the last two decades have led to the production of very high spatial resolution multispectral images. In order to adapt to this rapid development and handle these data, object-based analysis has emerged. A critical part of such an analysis is image segmentation. The selection of optimal segmentation parameters' values generates a qualitative segmentation output and has a direct impact on feature extraction and subsequent overall classification accuracy. Even though several image segmentation methods have been developed and suggested in the literature, each of them has advantages and disadvantages. This article presents the conceptual characteristics of image segmentation methods with a special focus on semantic segmentation. In addition, a meta-analysis was conducted through a comprehensive review of recent image segmentation case studies. It includes statistics and quantitative data regarding the applied segmentation algorithm, the software utilized and the data source among others. Since there is no miraculous segmentation algorithm, the statistical results depict only the recent trend. Finally, a few interesting subjects are addressed, including identification of current problems, image segmentation on non-traditional data and hot topics for future research.

Object-Based Comparison of Data-Driven and Physics-Driven Satellite Estimates of Extreme Rainfall.

The Global Precipitation Measurement (GPM) constellation of spaceborne sensors provides a variety of direct and indirect measurements of precipitation processes. Such observations can be employed to derive spatially and temporally consistent gridded precipitation estimates either via data-driven retrieval algorithms or by assimilation into physically based numerical weather models. We compare the data-driven Integrated Multisatellite Retrievals for GPM (IMERG) and the assimilation-enabled NASA-Unified Weather Research and Forecasting (NU-WRF) model against Stage IV reference precipitation for four major extreme rainfall events in the southeastern United States using an object-based analysis framework that decomposes gridded precipitation fields into storm objects. As an alternative to conventional "grid-by-grid analysis," the object-based approach provides a promising way to diagnose spatial properties of storms, trace them through space and time, and connect their accuracy to storm types and input data sources. The evolution of two tropical cyclones are generally captured by IMERG and NU-WRF, while the less organized spatial patterns of two mesoscale convective systems pose challenges for both. NU-WRF rain rates are generally more accurate, while IMERG better captures storm location and shape. Both show higher skill in detecting large, intense storms compared to smaller, weaker storms. IMERG's accuracy depends on the input microwave and infrared data sources; NU-WRF does not appear to exhibit this dependence. Findings highlight that an object-oriented view can provide deeper insights into satellite precipitation performance and that the satellite precipitation community should further explore the potential for "hybrid" data-driven and physics-driven estimates in order to make optimal usage of satellite observations.

Scale Optimization in Topographic and Hydrographic Feature Mapping Using Fractal Analysis

A new method for selecting optimal scales when mapping topographic or hydrographic features is introduced. The method employs rank-size partition of heavy-tailed distributions to detect nodes of rescaling invariance in the underlying hierarchy of the dataset. These nodes, known as head/tail breaks, can be used to indicate optimal scales. Then, the Fractal Net Evolution Assessment (FNEA) segmentation algorithm is applied with the topographic or hydrographic surfaces to produce optimally scaled objects. A topological transformation allows linking the two processes (partition and segmentation), while fractal dimension of the rescaling process is employed as an optimality metric. The new method is experimented with the two biggest river basins in Greece, namely Pinios and Acheloos river basins, using a digital elevation model as the only input dataset. The method proved successful in identifying a set of optimal scales for mapping elevation, slope, and flow accumulation. Deviation from the ideal conditions for implementing head/tail breaks are discussed. Implementation of the method requires an object-based analysis program and few common geospatial functions embedded in most GIS programs. The new method will assist in revealing underlying environmental processes in a variety of earth science fields and, thus, assist in land management decision-making and mapping generalization.

Quantifying highly dynamic urban landscapes: Integrating object-based image analysis with Landsat time series data

Urban landscapes are highly dynamic with changes frequently occurring at short time intervals. Although the Landsat data archive allows the use of high-density time-series data to quantify such dynamics, the approaches that can fully address the spatial and temporal complexity of the urban landscape are still lacking. A new approach is presented for accurately quantifying urban landscape dynamics. Information regarding when and where a change occurs, what type of change exists, and how often it happens are incorporated. The new approach integrates object-based image analysis and time-series change detection techniques by using all available Landsat images for several decades. This approach was tested on the rapidly urbanizing city of Shenzhen, China from 1986 to 2017. Land cover changes in both long- and short-time intervals can be proficiently detected with an overall accuracy of 90.65% and a user’s accuracy of 92.18% and 82.40% for “No change” and “Change”, respectively. The frequency and time of change can be explicitly displayed while incorporating the advantages of object-based image analysis and time-series change detection. The efficiency of the change analysis can be greatly increased because the object-based analysis greatly reduces the number of analyzed units. The new approach can accurately and efficiently detect the land cover change for quantifying urban landscape dynamics. Integrating the object and the remotely sensed time-series data has the potential to link the physical and socio-economic properties together for facilitating sustainable landscape planning.

Assessing the Impact of Stochastic Perturbations in Cloud Microphysics using GOES-16 Infrared Brightness Temperatures

AbstractIn this study, infrared brightness temperatures (BTs) are used to examine how applying stochastic perturbed parameter (SPP) methodology to the widely used Thompson–Eidhammer cloud microphysics scheme impacts the cloud field in high-resolution forecasts. Modifications are made to add stochastic perturbations to three parameters controlling cloud generation and dissipation processes. Two five-member ensembles are generated, one using the microphysics parameter perturbations (SPP-MP) and another where white noise perturbations were added to potential temperature fields at initialization time (Control). The impact of the SPP method was assessed using simulated and observed GOES-16 BTs. This analysis uses pixel-based and object-based methods to assess the impact on the cloud field. Pixel-based methods revealed that the SPP-MP BTs are slightly more accurate than the Control BTs. However, too few pixels with a BT lower than 270 K result in a positive bias compared to the observations. A negative bias compared to the observations is observed when only analyzing lower BTs. The spread of the ensemble BTs was analyzed using the continuous ranked probability score differences, with the SPP-MP ensemble BTs having less (more) spread during May (January) compared to the Control. Object-based analysis using the Method for Object-Based Diagnostic Evaluation revealed the upper-level cloud objects are smaller in the SPP-MP ensemble than the Control but a lower bias exists in the SPP-MP BTs compared to the Control BTs when overlapping matching objects. However, there is no clear distinction between the SPP-MP and Control ensemble members during the evolution of objects, Overall, the SPP-MP perturbations result in lower BTs compared to the Control ensemble and more cloudy pixels.

Evaluation of Rainfall Forecasts by Three Mesoscale Models during the Mei-Yu Season of 2008 in Taiwan. Part III: Application of an Object-Oriented Verification Method

In this study, the performances of Mei-yu (May–June) quantitative precipitation forecasts (QPFs) in Taiwan by three mesoscale models: the Cloud-Resolving Storm Simulator (CReSS), the Central Weather Bureau (CWB) Weather Research and Forecasting (WRF), and the CWB Non-hydrostatic Forecast System (NFS) are explored and compared using an newly-developed object-oriented verification method, with particular focus on the various properties or attributes of rainfall objects identified. Against a merged dataset from ~400 rain gauges in Taiwan and the Tropical Rainfall Measuring Mission (TRMM) data in the 2008 season, the object-based analysis is carried out to complement the subjective analysis in a parallel study. The Mei-yu QPF skill is seen to vary with different aspects of rainfall objects among the three models. The CReSS model has a total rainfall production closest to the observation but a large number of smaller objects, resulting in more frequent and concentrated rainfall. In contrast, both WRF and NFS tend to under-forecast the number of objects and total rainfall, but with a higher proportion of bigger objects. Location errors inferred from object centroid locations appear in all three models, as CReSS, NFS, and WRF exhibit a tendency to simulate objects slightly south, east, and northwest with respect to the observation. Most rainfall objects are aligned close to an E–W direction in CReSS, in best agreement with the observation, but many towards the NE–SW direction in both WRF and NFS. For each model, the objects are matched with the observed ones, and the results of the matched pairs are also discussed. Overall, though preliminarily, the CReSS model, with a finer grid size, emerges as best performing model for Mei-yu QPFs.

Object-Based Analyses in FIJI/ImageJ to Measure Local RNA Translation Sites in Neurites in Response to Aβ1-42 Oligomers.

Subcellular protein delivery is especially important in signal transduction and cell behavior, and is typically achieved by localization signals within the protein. However, protein delivery can also rely on localization of mRNAs that are translated at target sites. Although once considered heretical, RNA localization has proven to be highly conserved in eukaryotes. RNA localization and localized translation are especially relevant in polarized cells like neurons where neurites extend dozens to hundreds of centimeters away from the soma. Local translation confers dendrites and axons the capacity to respond to their environment in an acute manner without fully relying on somatic signals. The relevance of local protein synthesis in neuron development, maintenance and disease has not been fully acknowledged until recent years, partly due to the limited amount of locally produced proteins. For instance, in hippocampal neurons levels of newly synthesized somatic proteins can be more than 20–30 times greater than translation levels of neuritic proteins. Thus local translation events can be easily overlooked under the microscope. Here we describe an object-based analysis used to visualize and quantify local RNA translation sites in neurites. Newly synthesized proteins are tagged with puromycin and endogenous RNAs labeled with SYTO. After imaging, signals corresponding to neuritic RNAs and proteins are filtered with a Laplacian operator to enhance the edges. Resulting pixels are converted into objects and selected by automatic masking followed by signal smoothing. Objects corresponding to RNA or protein and colocalized objects (RNA and protein) are quantified along individual neurites. Colocalization between RNA and protein in neurites correspond to newly synthesized proteins arising from localized RNAs and represent localized translation sites. To test the validity of our analyses we have compared control neurons to Aβ1–42-treated neurons. Aβ is involved in the pathology of Alzheimer’s disease and was previously reported to induce local translation in axons and dendrites which in turn contributes to the disease. We have observed that Aβ increases the synthesis of neuritic proteins as well as the fraction of translating RNAs in distal sites of the neurite, suggesting an induction of local protein synthesis. Our results thus confirm previous reports and validate our quantification method.

Detailed Lacustrine Calving Iceberg Inventory from Very High Resolution Optical Imagery and Object-Based Image Analysis

In the field of iceberg and glacier calving studies, it is important to collect comprehensive datasets of populations of icebergs. Particularly, calving of lake-terminating glaciers has been understudied. The aim of this work is to present an object-based method of iceberg detection and to create an inventory of icebergs located in a proglacial lagoon of San Quintín glacier, Northern Patagonia Icefield, Chile. This dataset is created using high-resolution WorldView-2 imagery and a derived DEM. We use it to briefly discuss the iceberg size distribution and area–volume scaling. Segmentation of the multispectral imagery produced a map of objects, which were classified with use of Random Forest supervised classification algorithm. An intermediate classification product was corrected with a ruleset exploiting contextual information and a watershed algorithm that was used to divide multiple touching icebergs into separate objects. Common theoretical heavy-tail statistical distributions were tested as descriptors of iceberg area and volume distributions. Power law models were proposed for the area–volume relationship. The proposed method performed well over the open lake detecting correctly icebergs in all size bands except 5–15 m2 where many icebergs were missed. A section of the lagoon with ice melange was not reliably mapped due to uniformity of the area in the imagery and DEM. The precision of the DEM limited the scaling effort to icebergs taller than 1.7 m and larger than 99 m2, despite the inventory containing icebergs as small as 4 m2. The work demonstrates viability of object-based analysis for lacustrine iceberg detection and shows that the statistical properties of iceberg population at San Quintín glacier match those of populations produced by tidewater glaciers.

Object-Based Analysis Using Unmanned Aerial Vehicles (UAVs) for Site-Specific Landslide Assessment

The increased development of computer vision technology combined with the increased availability of innovative platforms with ultra-high-resolution sensors, has generated new opportunities and fields for investigation in the engineering geology domain in general and landslide identification and characterization in particular. During the last decade, the so-called Unmanned Aerial Vehicles (UAVs) have been evaluated for diverse applications such as 3D terrain analysis, slope stability, mass movement hazard and risk management. Their advantages of detailed data acquisition at a low cost and effective performance identifies them as leading platforms for site-specific 3D modelling. In this study, the proposed methodology has been developed based on Object-Based Image Analysis (OBIA) and fusion of multivariate data resulted from UAV photogrammetry processing in order to take full advantage of the produced data. Two landslide case studies within the territory of Greece, with different geological and geomorphological characteristics, have been investigated in order to assess the developed landslide detection and characterization algorithm performance in distinct scenarios. The methodology outputs demonstrate the potential for an accurate characterization of individual landslide objects within this natural process based on ultra high-resolution data from close range photogrammetry and OBIA techniques for landslide conceptualization. This proposed study shows that UAV-based landslide modelling on the specific case sites provides a detailed characterization of local scale events in an automated sense with high adaptability on the specific case site.

Tracking changes in soil organic carbon across the heterogeneous agricultural landscape of the Lower Fraser Valley of British Columbia

Increasing soil organic carbon (SOC) can improve the capacity of agricultural systems to both adapt to and mitigate climate change. Despite its importance, the current understanding of the magnitude or even the direction of SOC change in agricultural landscapes is limited. While changes in land use/land cover (LULC) and climate are among the main drivers of changes in SOC, their relative importance for the spatiotemporal assessment of SOC is unclear. This study evaluated LULC and SOC dynamics using archived and recent soil samples, remote sensing, and digital soil mapping in the Lower Fraser Valley of British Columbia, Canada. We combined both pixel- and object-based analysis of Landsat satellite imagery to assess LULC changes from 1984 to 2018. We achieved an overall accuracy of 81% and kappa coefficient of 0.77 for LULC classification using a random forest model. For predicting SOC for the same time period, we applied soil and vegetation indices derived from Landsat images, topographic indices, historic soil survey variables, and climate data in a random forest model. The SOC prediction of 2018 resulted in a coefficient of determination (R2) of 0.67, concordance correlation coefficient (CCC) of 0.76, and normalized root mean square error (nRMSE) of 0.12. For 1984, the SOC prediction accuracies were 0.46, 0.58, and 0.18 for R2, CCC, and nRMSE, respectively. We detected SOC loss in 61%, gain in 12%, while 27% remained unchanged across the study area. Although we detected large losses of SOC due to LULC change, the majority of the SOC losses across the landscape were attributed to areas that were remained in the same type of agricultural production since 1984. Climate variability did not, however, have a strong effect on SOC changes. These results can inform decision making in the study area to support sustainable LULC management for enhancing SOC sequestration.

Modeling strawberry biomass and leaf area using object-based analysis of high-resolution images

Quantifying canopy biophysical parameters is critical to agricultural research and farm management. In this study, strawberry dry biomass and leaf area were modeled statistically using high spatial and temporal resolution imagery. A mobile field data acquisition system was used to acquire thousands of very high resolution (~0.5 mm) close-range images seven times throughout the strawberry growing season. Ortho-mosaics and dense point clouds were generated through Structure from Motion (SfM) and used in Object-Based Image Analysis (OBIA) at the sub-leaf level to extract canopy structure variables such as planimetric canopy area, canopy average height, and canopy smoothness metric. Regression analysis was carried out using these image-derived canopy variables as predictors to model leaf area (R2 = 0.79; ten-fold cross-validation RMSE = 0.056 m2) and dry biomass (R2 = 0.84; ten-fold cross-validation RMSE = 7.72 g) obtained through destructive measurements. Results indicate consistent predictive power through the season and across 17 strawberry genotypes. The study showed that the canopy smoothness metric developed in this study as an indicator of canopy density could complement other variables (planimetric canopy area, canopy average height) that describe canopy geometric properties.