Search

The early stages of forest fires are usually accompanied by large amounts of smoke, so smoke identification and concentration inversion can effectively warn and locate the fire point. To achieve smoke identification and concentration inversion, this paper first analysed the spectral differences between smoke and underlying background features based on the samples selected from Landsat 8/OLI fire data of Qiqian forest in Inner Mongolia, China, andNew South Wales, Australia. Then, the sensitive band of smoke identification was screened and the relative smoke concentration indices were established. Finally, a mathematical model between the indices and smoke concentration level was constructed and the best inversion model was established by accuracy test based on samples in each background. The results show that (1) the reflectance increases with increasing smoke concentration in B1-B4 bands under different backgrounds (vegetation, water, bare ground), and the reflectance was significantly different between different concentrations of smoke; (2) the smoke concentration inversion model form is y = kx + b with R2 > 0.8 (x denotes the smoke concentration level, y denotes the relative smoke concentration index), and the visible brightness index (VBI = B1 + B2 + B3 + B4) was the best relative smoke concentration index under different background; (3) the model's accuracy was tested using each background smoke sample with a known level of smoke concentration, and the inversion accuracy was above 90%. The results show that the VBI can effectively reflect the spectral variation between different smoke concentrations and background features; the variation is closely related to smoke concentration which determines that the mathematical correlation model between VBI and relative smoke concentration can be established; the model is accuracy and can be used for the relative smoke concentration inversion of forest fire cases in different study areas. [ABSTRACT FROM AUTHOR]

Many intermodality models (IMMs) have been established to define how panchromatic (Pan) high-frequency details should be injected into the multispectral (MS) bands in pan-sharpening. In this paper, IMMs are categorized into two types, namely regression coefficient (RC) models and intensity modulation (IM) models. The RC is simply the ratio of the local covariance ofxandyto the local variance ofx, wherexandyare the approximation of the Pan image at the MS scale and the MS image, respectively. To compare these two IMM types, an ideal IMM representing the proper amount of high-frequency details which should be injected is formulated. This ideal IMM is exactly the same formulation as the RC model except thatxandyare high-resolution Pan and MS images, respectively. Therefore, compared to the ideal IMM, the RC model is influenced by the scale effect since the local variance and covariance change as functions of scale. The IM model, represented by the model in the high-pass filtering (HPM) method, may introduce improper injections which are affected by landscape complexity, spectral response function, atmospheric conditions, and offset values for converting radiance into digital number (DN). Using the information from the high-resolution Pan image, an improved HPM method is proposed for shadow regions, which are quite common in very high-resolution images. Experimental results using a WorldView-2 image covering an urban area and a QuickBird image covering a rural area confirmed our analytical conclusions on different IMMs and indicate that the improved HPM model performs better than the original HPM. [ABSTRACT FROM PUBLISHER]

A remote sensing approach is proposed in this paper for monitoring instantaneous underwater light climate in Dahuchi Lake, China. Water constituents that are optically significant to the lake water can be retrieved through this method based on an artificial neural network (ANN) algorithm with a satisfactory accuracy. In this study, the ANN was trained by spectral simulations which were generated by a water colour simulation model (WASI). An underwater light attenuation model has been built by fitting the field measurements and applied to calculate photosynthetically available radiation (PAR) at the bottom of the lake from the inputs of hyperspectral reflectance above the water surface and the water depth. Results show that this method can predict the concentration of suspended particle matter (SPM) with a high accuracy and explain more than half of the underwater PAR variation with a standard deviation of 94.8 μmol s-1 m-2. In order to test the portability of the algorithm, it has also been extended to three bands of Landsat TM image. Compared with hyperspectral inputs, ANN with broadband inputs has lower but still acceptable accuracy on SPM retrieval. The advantage of this method lies in its time-efficient and cost-efficient monitoring of underwater light climate, i.e. the simulations generated by a semi-physical model can be used for ANN training, instead of using a large number of field measurements. [ABSTRACT FROM AUTHOR]

The use of land cover change maps is subject to the propagation of errors involved in classifying multi-temporal land cover maps. Understanding the link between classification processes and error propagation helps to determine appropriate classification models to mitigate the error propagation rate. In this paper, we present a simulation analysis on error propagation in land cover change detection using three classification models: a non-contextual model, a contextual model based on spatial smoothing, and a contextual model based on Markov random fields (MRF). A spatial simulation approach based on simulated annealing was developed with careful experimental designs to control two related factors including the spatial/temporal patterns of estimation errors associated with spectral probabilities. The results showed that the contextual classification model based on MRF had the smallest error propagation rate while the non-contextual classification model had the largest rate under all scenarios. The two factors had different effects on the error propagation for different classification models. For the non-contextual model, increasing temporal correlation of errors could reduce the error propagation rate while spatial autocorrelation of errors did not have a big impact on the error propagation. For the two contextual classification models, the use of contextual information significantly reduced the error propagation rate. However, the value of contextual information in mitigating error propagation was highly dependent on the spatial autocorrelation of the errors. The impact of the temporal correlation of errors was weakened in the contextual models. [ABSTRACT FROM AUTHOR]

The present paper develops three predictive models for daily total ozone concentration over Arosa, Switzerland. The models are artificial neural network, multiple linear regression, and persistence forecast. Each model was judged for their predictive ability using analysis of variance, Pearson correlation study, and scatterplot analysis. Prediction errors were computed for each model. After painstaking analysis it was established that artificial neural network produces better forecasts than the statistical approaches like multiple linear regression and persistence forecast models. [ABSTRACT FROM AUTHOR]

Mass movements (MM) represent a serious threat to human life and activities in most mountainous areas. However, due to the rugged nature of such terrain, it is often difficult to detect such phenomena in remote areas. Hence, satellite imagery offers many attractions for the examination of MM in such environments, especially in less developed nations in which resources are stretched and levels of environmental information limited. There is a need to ensure that the techniques and images used are effective, reliable, and cheap in terms of the amount and accuracy of data that can be extracted. Taking Lebanon as a case study, this paper compares the applicability of different satellite data sensors (Landsat TM (Thematic Mapper), IRS (Indian Remote Sensing Satellite), SPOT4 (Système Probatoire pour l'Observation de la Terre)) and preferred image-processing techniques (False Colour Composite 'FCC', pan-sharpen, principal-component analysis 'PCA', Anaglyph) for the mapping of MM recognized as landslides, rock and debris falls, and earth flows. Results from the imagery have been validated by field surveys and analysis of IKONOS imagery acquired in some locations witnessing major MM during long periods. Then, levels of accuracies of detected MM from satellite imageries were plotted. This study has demonstrated that the anaglyph produced from the two panchromatic stereo-pairs SPOT4 images remains the most effective tool setting the needed 3D properties for visual interpretation and showing a maximum accuracy level of 67%. The PCA pan-sharpened Landsat TM-IRS image gave better results in detecting MM, among other processing techniques, with a maximum accuracy level of 62%. [ABSTRACT FROM AUTHOR]

Watershed transformation in mathematical morphology is a powerful morphological tool for image segmentation that is usually defined for greyscale images and applied to the gradient magnitude of an image. This paper presents an extension of the watershed algorithm for multispectral image segmentation. A vector-based morphological approach is proposed to compute gradient magnitude from multispectral imagery, which is then input into watershed transformation for image segmentation. The gradient magnitude is obtained at multiple scales. After an automatic elimination of local irrelevant minima, a watershed transformation is applied to segment the image. The segmentation results were evaluated and compared with other multispectral image segmentation methods, in terms of visual inspection, and object-based image classification using high resolution multispectral images. The experimental results indicate that the proposed method can produce accurate segmentation results and higher classification accuracy, if the scales and contrast parameter are appropriately selected in the gradient computation and subsequent local minima elimination. The proposed method shows encouraging results and can be used for segmentation of high resolution multispectral imagery and object based classification. [ABSTRACT FROM AUTHOR]

This paper presents the results of an airborne thermal infrared (TIR) experiment. The data were obtained during 6-9 February 1992 period in the Bay of Bothnia, the northern section of the Baltic Sea, in connection with an ERS-1 field campaign. Two Aegema model 880 TIR cameras were used with nominal wavelengths at 5 and 10 µm, attached to the front of the helicopter with a Tyler mount. The camera's thermal resolution is 0.2 K; after corrections for atmosphere effects, the surface temperature accuracy is 0.5 K. The spatial resolution of individual images is 70 cm at the 300-m flight altitude. The measurement programme was successful, producing high-quality TIR data over ice for two different days, even under difficult weather conditions. Ice temperatures ranged from open water temperatures to 261 K for fast ice conditions on the day on the first flight. The standard deviation of the surface temperature, generally, increased with ice thickness with a value of ∼0.4 K for maximum thickness, and autocorrelation length scales not exceeding a value of 5 m. Generally, all the higher values of standard deviation (>0.7) of surface temperatures were for scenes with mixed ice/open water. The results show that TIR has a substantial ability to classify ice type and thickness when the air temperature is less than 269 K, from open water at the freezing point to thin nilas and thicknesses up to 20-45 cm in the fast ice zone. In addition, a quasi-steady sea-ice model is used to provide a physical interpretation of the sea-ice surface temperatures. The use of the model requires information on the atmospheric surface layer and snow thickness data, together with calibration points. The model worked well when the air temperature was around 260 K or less. [ABSTRACT FROM AUTHOR]

This paper introduces a methodology able to discriminate between non-stressed plants and N, P and K stress symptoms in spring barley grown under controlled conditions, utilizing the spectral and spatial dimensions simultaneously. Nine spectral measurements in the range 450-1000 nm were taken for each plant. The measuring points were spatially located at the tip, middle and base of the last three fully developed leaves. This design generated a four-way data set consisting of measurements as a function of (i) the specific plant, (ii) the spectral wavelength, (iii) the plant leaf position and (iv) the position on the leaf. Multiway partial least squares regression analysis with dummy variables was able correctly to classify the four nutrient conditions with 92% accuracy regardless of the respective growth stages within a time window of 2 weeks. The addition of the spatial dimension to the spectral dimension proved to be a promising nutrient diagnostic tool. Without performance loss it was possible to reduce the hyperspectral resolution to a resolution of three wavelengths. The three selected 2 nm wide bands were R450, R700 and R810, which agrees well with the literature on plant spectral reflectance in relation to nutritional stress. [ABSTRACT FROM AUTHOR]

In this study a very high resolution image from the QuickBird satellite was used to detect new greenhouses built since the last update of the information system utilized for the study area. The area, located in the southeast of Spain, has the highest concentration of greenhouses in Europe, which makes it the heart of the economy of this region. The methodology proposed in this paper is based on the comparison of the classification of a current image with the information system corresponding to the last update. Maximum likelihood classification method was employed and different band combinations were used to define the training areas and to carry out the classification process. The optimal band combination for the detection of greenhouses was calculated by means of a variance analysis. The process was completed with the delineation of the new greenhouses with two algorithms programmed in Visual Basic 6.0, one to eliminate the loops shown around the greenhouses detected, and the other one, based on the Hough transformation, to delineate the contour of the polygons corresponding to the new greenhouses. The proposed methodology achieved (1) a value for true greenhouse surface of about 91.45% of the whole surface, (2) a very low value for undetected greenhouses (five greenhouses from a total of 202 that were built, representing 1.49% of the surface of new greenhouses), and (3) a low number of pixels wrongly classified as greenhouses. [ABSTRACT FROM AUTHOR]

The emergence of high temporal resolution satellite data such as MODIS enables timely monitoring of locust outbreaks from space. This monitoring is hampered by the effect of random atmospheric variations on satellite imagery, which may be suppressed through temporal filtering. This paper aims to evaluate the utility of temporally filtering successive MODIS data in monitoring an outbreak in East China. Of the eight vegetation indices examined, the commonly used NDVI was the most indicative of varying vegetation conditions caused by locust infestation inside the study area. The averaging of three successive days of satellite data improves the R 2 value of NDVI regression models by 0.227 over single-day data. It also outperforms the data averaged from two successive days (a broader window size was not attempted due to the short span of the study period). Temporally, NDVI changed at varying rates daily during the outbreak. Early in the outbreak it increased at a reduced pace until 7.5 days. Afterwards it started to decrease at an accelerated rate. If temporally filtered with a proper window size, successive MODIS data allow the outbreak to be monitored accurately ( R 2 = 0.696). [ABSTRACT FROM AUTHOR]

In this paper, the rotational transformation process is explained as a problem of rotation of remotely sensed data in the variance-covariance space. In particular, the rotation which maximizes the covariance-variance ratio is examined in detail for various land use and land cover classes in the Bombay suburban and Thana district area in India. A statistical approach to determine transformed components and other statistical variables on different band combinations is discussed. The results are analysed, and the best possible combinations selected for accurate classification are presented. [ABSTRACT FROM AUTHOR]