Eufemia Tarantino

Performance evaluation of object based greenhouse detection from Sentinel-2 MSI and Landsat 8 OLI data: A case study from Almería (Spain)

Abstract This paper shows the first comparison between data from Sentinel-2 (S2) Multi Spectral Instrument (MSI) and Landsat 8 (L8) Operational Land Imager (OLI) headed up to greenhouse detection. Two closely related in time scenes, one for each sensor, were classified by using Object Based Image Analysis and Random Forest (RF). The RF input consisted of several object-based features computed from spectral bands and including mean values, spectral indices and textural features. S2 and L8 data comparisons were also extended using a common segmentation dataset extracted form VHR World-View 2 (WV2) imagery to test differences only due to their specific spectral contribution. The best band combinations to perform segmentation were found through a modified version of the Euclidian Distance 2 index. Four different RF classifications schemes were considered achieving 89.1%, 91.3%, 90.9% and 93.4% as the best overall accuracies respectively, evaluated over the whole study area.

Mapping Rural Areas with Widespread Plastic Covered Vineyards Using True Color Aerial Data

Plastic covering is used worldwide to protect crops against damaging growing conditions. This agricultural practice raises some controversial issues. While it significantly impacts on local economic vitality, plasticulture also shows several environmental affects. In the Apulia Region (Italy) the wide-spreading of artificial plastic coverings for vineyard protection has showed negative consequences on the hydrogeological balance of soils as well as on the visual quality of rural landscape. In order to monitor and manage this phenomenon, a detailed site mapping has become essential. In this study an efficient object-based classification procedure from Very High Spatial Resolution (VHSR) true color aerial data was developed on eight test areas located in the Ionian area of the Apulia Region in order to support the updating of the existing land use database aimed at plastic covered vineyard monitoring.

A data fusion algorithm based on the Kalman filter to estimate leaf area index evolution in durum wheat by using field measurements and MODIS surface reflectance data

ABSTRACT The use of leaf area index (LAI) is essential in ecosystem and agronomic studies since it measures energy and gas exchanges between vegetation and atmosphere. In the last decades, LAI values have widely been estimated from passive remotely sensed data although estimated results were often affected by noise and measurement uncertainties. In this article, we propose a Kalman filter algorithm in order to estimate the time evolution of LAI by combining field-measured and Moderate Resolution Imaging Spectroradiometer (MODIS) surface reflectance data. The scalar equation of the dynamic LAI model (state transition model) was derived by the field-measured data while the MODIS red, near-infrared and shortwave infrared reflectance data were used to implement the observation model. The reflectance data were linked to LAI by using the reduced simple ratio. The method was tested in an experimental field located in the north-western part of the Apulia region (Italy). The results showed a good agreement between the LAI estimated through the algorithm and the LAI derived from field data, with a coefficient of determination (R2) of 0.96 and a corresponding root mean square error of 0.124.

A tree counting algorithm for precision agriculture tasks

Abstract This study proposes an automatic procedure for individual fruit tree identification using GeoEye-1 sensor data. Depending on site-specific pruning practices, the morphologic characteristics of tree crowns may generate one or more brightness peaks (tree top) on the imagery. To optimize tree counting and to minimize typical background noises from orchards (i.e. bare soil, weeds, and man-made objects), a four-step algorithm was implemented with spatial transforms and functions suitable for spaced stands (asymmetrical smoothing filter, local minimum filter, mask layer, and spatial aggregation operator). System performance was evaluated through objective criteria, showing consistent results in fast capturing tree position for precision agriculture tasks.

Semi-automatic detection of linear archaeological traces from orthorectified aerial images

Abstract This paper presents a semi-automatic approach for archaeological traces detection from aerial images. The method developed was based on the multiphase active contour model (ACM). The image was segmented into three competing regions to improve the visibility of buried remains showing in the image as crop marks (i.e. centuriations, agricultural allocations, ancient roads, etc.). An initial determination of relevant traces can be quickly carried out by the operator by sketching straight lines close to the traces. Subsequently, tuning parameters (i.e. eccentricity, orientation, minimum area and distance from input line) are used to remove non-target objects and parameterize the detected traces. The algorithm and graphical user interface for this method were developed in a MATLAB environment and tested on high resolution orthorectified aerial images. A qualitative analysis of the method was lastly performed by comparing the traces extracted with ancient traces verified by archaeologists.

Testing high spatial resolution WorldView-2 imagery for retrieving the leaf area index

This work analyzes the potentiality of WorldView-2 satellite data for retrieving the Leaf Area Index (LAI) area located in Apulia, the most Eastern region of Italy, overlooking the Adriatic and Ionian seas. Lacking contemporary in-situ measurements, the semi-empiric method of Clevers (1989) (CLAIR model) was chosen as a feasible image-based LAI retrieval method, which is based on an inverse exponential relationship between the LAI and the WDVI (Weighted Difference Vegetation Index) with relation to different land covers. Results were examined in homogeneous land cover classes and compared with values obtained in recent literature.

Combining ad hoc spectral indices based on LANDSAT-8 OLI/TIRS sensor data for the detection of plastic cover vineyard

In this article, we are proposing a method using Landsat-8 Operational Land Imager and Thermal Infrared Sensor data for agricultural plastic cover detection. Four normalized difference indices were combined in the procedure described to achieve consistent results: the green Normalized Difference Vegetation Index and three ad hoc spectral indices purposely created for this study (rescaled brightness temperature, Plastic Surface Index and Normalized Difference Sandy Index). The sampling time related to the preliminary collection of spectral information on plastic surfaces was reduced using information gathered through the Quality Assessment and Cloud Quality bands. The overall accuracies observed were on average higher than 80%,and the low cost of the open data set used, lacking ancillary data, demonstrated the reliability of the proposed method, proving its suitability for environmental and agricultural monitoring over large areas.

Extracting Buildings from True Color Stereo Aerial Images Using a Decision Making Strategy

The automatic extraction of buildings from true color stereo aerial imagery in a dense built-up area is the main focus of this paper. Our approach strategy aimed at reducing the complexity of the image content by means of a three-step procedure combining reliable geospatial image analysis techniques. Even if it is a rudimentary first step towards a more general approach, the method presented proved useful in urban sprawl studies for rapid map production in flat area by retrieving indispensable information on buildings from scanned historic aerial photography. After the preliminary creation of a photogrammetric model to manage Digital Surface Model and orthophotos, five intermediate mask-layers data (Elevation, Slope, Vegetation, Shadow, Canny, Shadow, Edges) were processed through the combined use of remote sensing image processing and GIS software environments. Lastly, a rectangular building block model without roof structures (Level of Detail, LoD1) was automatically generated. System performance was evaluated with objective criteria, showing good results in a complex urban area featuring various types of building objects.

A Class-Oriented Strategy for Features Extraction from Multidate ASTER Imagery

In this paper we propose a hybrid classification method, adopting the best features extraction strategy for each land cover class on multidate ASTER data. To enable an effective comparison among images, Multivariate Alteration Detection (MAD) transformation was applied in the pre-processing phase, because of its high level of automation and reliability in the enhancement of change information among different images. Consequently, different features identification procedures, both spectral and object-based, were implemented to overcome problems of misclassification among classes with similar spectral response. Lastly, a post-classification comparison was performed on multidate ASTER-derived land cover (LC) maps to evaluate the effects of change in the study area.

Steerable filtering in interactive tracing of archaeological linear features using digital true colour aerial images

In this study, a semi-automatic approach to support archaeological line tracing is proposed. The suggested procedure is based on colour and texture information derived from orthorectified RGB digital aerial data and consists of four steps: (1) line sketching; (2) steerable filtering; (3) objects selection; and (4) straight line fitting and vectorisation. Good results were observed by evaluating the algorithm according to trace visibility and integrity, global difficulty and level of feature extraction. Further reliability tests were performed to study poor data initialisation and different annual seasonal land use at the same site.