Analysis of classification accuracy for pre-filtered multichannel remote sensing data

Noise is one of the main factors degrading the quality of original multichannel remote sensing data and its presence influences classification efficiency, object detection, etc. Thus, pre-filtering is often used to remove noise and improve the solving of final tasks of multichannel remote sensing. Recent studies indicate that a classical model of additive noise is not adequate enough for images formed by modern multichannel sensors operating in visible and infrared bands. However, this fact is often ignored by researchers designing noise removal methods and algorithms. Because of this, we focus on the classification of multichannel remote sensing images in the case of signal-dependent noise present in component images. Three approaches to filtering of multichannel images for the considered noise model are analysed, all based on discrete cosine transform in blocks. The study is carried out not only in terms of conventional efficiency metrics used in filtering (MSE) but also in terms of multichannel data classification accuracy (probability of correct classification, confusion matrix). The proposed classification system combines the pre-processing stage where a DCT-based filter processes the blocks of the multichannel remote sensing image and the classification stage. Two modern classifiers are employed, radial basis function neural network and support vector machines. Simulations are carried out for three-channel image of Landsat TM sensor. Different cases of learning are considered: using noise-free samples of the test multichannel image, the noisy multichannel image and the pre-filtered one. It is shown that the use of the pre-filtered image for training produces better classification in comparison to the case of learning for the noisy image. It is demonstrated that the best results for both groups of quantitative criteria are provided if a proposed 3D discrete cosine transform filter equipped by variance stabilizing transform is applied. The classification results obtained for data pre-filtered in different ways are in agreement for both considered classifiers. Comparison of classifier performance is carried out as well. The radial basis neural network classifier is less sensitive to noise in original images, but after pre-filtering the performance of both classifiers is approximately the same.

Coccolithophore surface distributions in the North Atlantic and their modulation of the air-sea flux of CO2 from 10 years of satellite Earth observation data

Coccolithophores are the primary oceanic phytoplankton responsible for the production of calcium carbonate (CaCO3). These climatically important plankton play a key role in the oceanic carbon cycle as a major contributor of carbon to the open ocean carbonate pump (similar to 50 %) and their calcification can affect the atmosphere-to-ocean (air-sea) uptake of carbon dioxide (CO2) through increasing the seawater partial pressure of CO2 (pCO(2)). Here we document variations in the areal extent of surface blooms of the globally important coccolithophore, Emiliania huxleyi, in the North Atlantic over a 10-year period (1998-2007), using Earth observation data from the Sea-viewing Wide Field-of-view Sensor (SeaWiFS). We calculate the annual mean sea surface areal coverage of E. huxleyi in the North Atlantic to be 474 000 +/- 104 000 km(2), which results in a net CaCO3 carbon (CaCO3-C) production of 0.14-1.71 Tg CaCO3-C per year. However, this surface coverage (and, thus, net production) can fluctuate inter-annually by -54/+81% about the mean value and is strongly correlated with the El Nino/Southern Oscillation (ENSO) climate oscillation index (r = 0.75, p < 0.02). Our analysis evaluates the spatial extent over which the E. huxleyi blooms in the North Atlantic can increase the pCO(2) and, thus, decrease the localised air-sea flux of atmospheric CO2. In regions where the blooms are prevalent, the average reduction in the monthly air-sea CO2 flux can reach 55%. The maximum reduction of the monthly air-sea CO2 flux in the time series is 155 %. This work suggests that the high variability, frequency and distribution of these calcifying plankton and their impact on pCO(2) should be considered if we are to fully understand the variability of the North Atlantic air-to-sea flux of CO2. We estimate that these blooms can reduce the annual N. Atlantic net sink atmospheric CO2 by between 3-28 %.

Coccolithophore surface distributions in the North Atlantic and their modulation of the air-sea flux of CO2 from 10 years of satellite Earth observation data

Coccolithophores are the primary oceanic phytoplankton responsible for the production of calcium carbonate (CaCO3). These climatically important plankton play a key role in the oceanic carbon cycle as a major contributor of carbon to the open ocean 5 carbonate pump (�50%) and their formation can affect the atmosphere-to-ocean (airsea) uptake of carbon dioxide (CO2) through increasing the seawater partial pressure of CO2 (pCO2). Here we document variations in the areal extent of surface blooms of the globally important coccolithophore, Emiliania huxleyi, in the North Atlantic over a 10-year period (1998–2007), using Earth observation data from the Sea-viewing Wide 10 Field of view Sensor (SeaWiFS).We calculate the annual mean surface areal coverage of E. huxleyi in the North Atlantic to be 474 000±119 000km2 yr−1, which results in a net CaCO3 production of 0.62±0.15 Tg CaCO3 carbon per year. However, this surface coverage and net production can fluctuate by −54/+81% about these mean values and are strongly correlated with the El Ni˜no/Southern Oscillation (ENSO) climate os15 cillation index (r =0.75, p<0.02). Our analysis evaluates the spatial extent over which the E. huxleyi blooms in the North Atlantic can increase the pCO2 and thus decrease the localised sink of atmospheric CO2. In regions where the blooms are prevalent, the average reduction in the monthly CO2 sink can reach 12 %. The maximum reduction of the monthly CO2 sink in the time series is 32 %. This work suggests that the high 20 variability, frequency and distribution of these calcifying plankton and their impact on pCO2 should be considered within modelling studies of the North Atlantic if we are to fully understand the variability of its air-to-sea CO2 flux.

Data processing of remotely sensed airborne hyperspectral data using the Airborne Processing Library (APL): Geocorrection algorithm descriptions and spatial accuracy assessment

Remote sensing airborne hyperspectral data are routinely used for applications including algorithm development for satellite sensors, environmental monitoring and atmospheric studies. Single flight lines of airborne hyperspectral data are often in the region of tens of gigabytes in size. This means that a single aircraft can collect terabytes of remotely sensed hyperspectral data during a single year. Before these data can be used for scientific analyses, they need to be radiometrically calibrated, synchronised with the aircraft's position and attitude and then geocorrected. To enable efficient processing of these large datasets the UK Airborne Research and Survey Facility has recently developed a software suite, the Airborne Processing Library (APL), for processing airborne hyperspectral data acquired from the Specim AISA Eagle and Hawk instruments. The APL toolbox allows users to radiometrically calibrate, geocorrect, reproject and resample airborne data. Each stage of the toolbox outputs data in the common Band Interleaved Lines (BILs) format, which allows its integration with other standard remote sensing software packages. APL was developed to be user-friendly and suitable for use on a workstation PC as well as for the automated processing of the facility; to this end APL can be used under both Windows and Linux environments on a single desktop machine or through a Grid engine. A graphical user interface also exists. In this paper we describe the Airborne Processing Library software, its algorithms and approach. We present example results from using APL with an AISA Eagle sensor and we assess its spatial accuracy using data from multiple flight lines collected during a campaign in 2008 together with in situ surveyed ground control points.

Composite front maps for improved visibility of dynamic sea-surface features on cloudy SeaWiFS and AVHRR data

Novel techniques have been developed for increasing the value of cloud-affected sequences of Advanced Very High Resolution Radiometer (AVHRR) sea-surface temperature (SST) data and Sea-viewing Wide Field-of-view Sensor (SeaWiFS) ocean colour data for visualising dynamic physical and biological oceanic processes such as fronts, eddies and blooms. The proposed composite front map approach is to combine the location, strength and persistence of all fronts observed over several days into a single map, which allows intuitive interpretation of mesoscale structures. This method achieves a synoptic view without blurring dynamic features, an inherent problem with conventional time-averaging compositing methods. Objective validation confirms a significant improvement in feature visibility on composite maps compared to individual front maps. A further novel aspect is the automated detection of ocean colour fronts, correctly locating 96% of chlorophyll fronts in a test data set. A sizeable data set of 13,000 AVHRR and 1200 SeaWiFS scenes automatically processed using this technique is applied to the study of dynamic processes off the Iberian Peninsula such as mesoscale eddy generation, and many additional applications are identified. Front map animations provide a unique insight into the evolution of upwelling and eddies.

Ocean Virtual Laboratory: A new way to explore multi-sensor synergy demonstrated over the Agulhas region

Ocean Virtual Laboratory is an ESA-funded project to prototype the concept of a single point of access for all satellite remote-sensing data with ancillary model output and in situ measurements for a given region. The idea is to provide easy access for the non-specialist to both data and state-of-the-art processing techniques and enable their easy analysis and display. The project, led by OceanDataLab, is being trialled in the region of the Agulhas Current, as it contains signals of strong contrast (due to very energetic upper ocean dynamics) and special SAR data acquisitions have been recorded there. The project also encourages the take up of Earth Observation data by developing training material to help those not in large scientific or governmental organizations make the best use of what data are available. The website for access is: http://ovl-project.oceandatalab.com/

Data processing of remotely sensed airborne hyperspectral data using the Airborne Processing Library (APL): Geocorrection algorithm descriptions and spatial accuracy assessment

Remote sensing airborne hyperspectral data are routinely used for applications including algorithm development for satellite sensors, environmental monitoring and atmospheric studies. Single flight lines of airborne hyperspectral data are often in the region of tens of gigabytes in size. This means that a single aircraft can collect terabytes of remotely sensed hyperspectral data during a single year. Before these data can be used for scientific analyses, they need to be radiometrically calibrated, synchronised with the aircraft's position and attitude and then geocorrected. To enable efficient processing of these large datasets the UK Airborne Research and Survey Facility has recently developed a software suite, the Airborne Processing Library (APL), for processing airborne hyperspectral data acquired from the Specim AISA Eagle and Hawk instruments. The APL toolbox allows users to radiometrically calibrate, geocorrect, reproject and resample airborne data. Each stage of the toolbox outputs data in the common Band Interleaved Lines (BILs) format, which allows its integration with other standard remote sensing software packages. APL was developed to be user-friendly and suitable for use on a workstation PC as well as for the automated processing of the facility; to this end APL can be used under both Windows and Linux environments on a single desktop machine or through a Grid engine. A graphical user interface also exists. In this paper we describe the Airborne Processing Library software, its algorithms and approach. We present example results from using APL with an AISA Eagle sensor and we assess its spatial accuracy using data from multiple flight lines collected during a campaign in 2008 together with in situ surveyed ground control points.