Augmenting LANDSAT MSS Data with Topographic Information for Enhanced Registration And Classification

This paper deals with problems arising in the classification of LANDSAT MSS data from rugged terrain. A digital terrain model (DTM) was found to be useful in several ways. For registration by cross-correlation, mountain ridges were extracted from both a synthetic image based on the DTM and a LANDSAT image. Information from the DTM, from thematic maps, and meteorological data were all used as ancillary data to aid in rapid snow cover determination without direct ground control in a large catchment area. In addition it is shown that the use of the DTM not only allows the assessment of relative and absolute snow distribution within given elevation zones, but also permits the extrapolation of snow cover into areas partly covered with clouds.     

Application of neural networks to radar image classification

A number of methods have been developed to classify ground terrain types from fully polarimetric synthetic aperture radar (SAR) images, and these techniques are often grouped into supervised and unsupervised approaches. Supervised methods have yielded higher accuracy than unsupervised techniques, but suffer from the need for human interaction to determine classes and training regions. In contrast, unsupervised methods determine classes automatically, but generally show limited ability to accurately divide terrain into natural classes. In this paper, a new terrain classification technique is introduced to determine terrain classes in polarimetric SAR images, utilizing unsupervised neural networks to provide automatic classification, and employing an iterative algorithm to improve the performance. Several types of unsupervised neural networks are first applied to the classification of SAR images, and the results are compared to those of more conventional unsupervised methods. Results show that one neural network method-Learning Vector Quantization (LVQ)-outperforms the conventional unsupervised classifiers, but is still inferior to supervised methods. To overcome this poor accuracy, an iterative algorithm is proposed where the SAR image is reclassified using a maximum likelihood (ML) classifier. It is shown that this algorithm converges, and significantly improves classification accuracy     

An Application of Canonical Correlation Analysis Technique to Land Cover Classification of LANDSAT Images

This research is an attempt to obtain more accurate land cover information from LANDSAT images. Canonical correlation analysis, which has not been widely used in the image classification community, was applied to the classification of a LANDSAT image. It was found that it is easy to select training areas on the classification using canonical correlation analysis in comparison with the maximum likelihood classifier of ERDAS^® software. In other words, the selected positions of training areas hardly affect the classification results using canonical correlation analysis. When the same training areas are used, the mapping accuracy of the canonical correlation classification results compared with the ground truth data is not lower than that of the maximum likelihood classifier. The kappa analysis for the canonical correlation classifier and the maximum likelihood classifier showed that the two methods are alike in classification accuracy. However, the canonical correlation classifier has better points than the maximum likelihood classifier in classification characteristics. Therefore, the classification using canonical correlation analysis applied in this research is effective for the extraction of land cover information from LANDSAT images and will be able to be put to practical use.     

Modeling the Radiant Transfers of Sparse Vegetation Canopies

The three-dimensional radiative transfer model of Kimes [2] was used to extend our understanding of the physical principles causing the scattering dynamics in sparse vegetation canopies (? 50-percent ground cover). The model was upgraded by including an aniotropic scattering algorithm for soil developed by Walthall et al. [7]. The model was validated using measured directional reflectance data that covered the entire exitance hemisphere. Two canopies were chosen to present in this study-an orchard grass canopy (50-percent ground cover) and a hard wheat canopy (11-percent ground cover). These canopies showed the typical scattering behavior of canopies with low and intermediate vegetation density. A red wavelength (0.58-0.68 pm) band was used throughout the study. A number of phenomena contributed to the directional reflectance distributions observed in the field. These include: 1) the strong anisotropic scattering properties of the soil, 2) the geometric effect of the vegetation probability of gap function on the soil anisotropy and solar irradiance, and 3) the anisotropic scattering of vegetation which is controlled by the phase function (for an infinitely small volume of representative leaves) and geometric Effect 1 (cause by layering of leaves). These phenomena as identified in this paper account for the major scattering behavior of observed data sets of directional reflectance distributions. Such knowledge provides an intelligent basis for defining specifications of earth-observing sensor systems and for inferring important aspects of physical and biological processes of the plant system.     

Crop Classification Using Airborne Radar and Landsat Data

Airborne radar data acquired with a 13.3-GHz scatterometer over a test site near Colby, KS, were used to investigate the statistical properties of the scattering coefficient of three types of vegetation cover and of bare soil. A statistical model for radar data was developed that incorporates signal fading and natural within-field variabilities. Estimates of the within-field and between-field coefficients of variation were obtained for each cover type and compared with similar quantities derived from Landsat images of the same fields. The second phase of this study consisted of evaluating the classification accuracy provided by Landsat alone, radar alone, and both sensors combined. The results indicate that the addition of radar to Landsat improves the classification accuracy by about 10 percentage points when the classification is performed on a pixel basis and by about 15 points when performed on a field-average basis. As with all crop-classification studies, these results pertain to the specific dates, geographic region, and crop categories.     

Genetic Sequence Classification and its Application to Cross-Species Homology Detection

Although large-scale classification studies of genetic sequence data are in progress around the world, very few studies compare different classification approaches, e.g. unsupervised and supervised, in terms of objective criteria such as classification accuracy and computational complexity. In this paper, we study such criteria for both unsupervised and supervised classification of a relatively large sequence data set. The unsupervised approach involves use of different sequence alignment algorithms (e.g., Smith-Waterman, FASTA and BLAST) followed by clustering using the Maximin algorithm. The supervised approach uses a suitable numeric encoding (relative frequencies of tuples of nucleotides followed by principal component analysis) which is fed to a Multi-layer Backpropagation Neural Network. Classification experiments conducted on IBM-SP parallel computers show that FASTA with unsupervised Maximin leads to best trade-off between accuracy and speed among all methods, followed by supervised neural networks as the second best approach. Finally, the different classifiers are applied to the problem of cross-species homology detection.     

Estimation and monitoring of bare soil/vegetation ratio with SPOT VEGETATION and HRVIR

Covering soils with vegetation during the fallow and planting seasons is one of the main ways to reduce water pollution, by restricting pollutant fluxes to aquatic systems. The bare soil/vegetation ratio monitoring can be carried out daily with a coarse spatial resolution using SPOT VEGETATION (1 km). Nevertheless, land-cover changes detected at a regional scale with this ratio may be due to winter vegetation cover changes as well as the influence of climatic events. Therefore, observed changes have to be validated from a local-scale analysis with higher spatial resolution data. The aim of this study is to develop a technique that allows high or low variations detected at a regional scale to be assessed from SPOT VEGETATION images with data acquired at a higher scale, SPOT High Resolution Visible and Infrared images in our case. In this study, the link between the images from the two sensors is achieved from the design of an artificial neural network method based on a Kohonen self-organizing map. The originality of this method lies in the use of external knowledge from ground observations and the use of temporal behavior to solve such a change of scale. Results of testing this method by using a potential change map based on the last few years' land-cover observations have shown a good correspondence between the observed and predicted bare soil/vegetation balance with regards to the spatial resolution difference between the two sensors.     

Automatic genre classification of music content: a survey

This paper reviews the state-of-the-art in automatic genre classification of music collections through three main paradigms: expert systems, unsupervised classification, and supervised classification. The paper discusses the importance of music genres with their definitions and hierarchies. It also presents techniques to extract meaningful information from audio data to characterize musical excerpts. The paper also presents the results of new emerging research fields and techniques that investigate the proximity of music genres     

Retrieval of vegetation parameters with SAR interferometry

The potential of SAR interferometric techniques for the retrieval of vegetation parameters was investigated using ERS-1 data over agricultural and forested test sites. In a first experiment an interferometrically derived forest map was generated. The classification was based on the interferometric correlation and the backscatter intensities. The result was geocoded, using the interferometrically derived height map generated from the same ERS SAR data pair, and validated with a conventional digital forest map. Forest mapping accuracies of around 90% and better were achieved. In a second experiment, multitemporal data over an agricultural site were used to investigate the potential of repeat-pass interferometry to monitor farming activity, crop development, and soil moisture variations. The interferometric correlation was used as an indicator of dense vegetation and geometric change. It was possible, for example, to identify harvesting by the high correlation of the post-harvest bare or stubble field. Decreasing interferometric correlation was observed as a consequence of crop growth     

Application of Dempster-Shafer evidence theory to unsupervisedclassification in multisource remote sensing

The aim of this paper is to show that Dempster-Shafer evidence theory may be successfully applied to unsupervised classification in multisource remote sensing. Dempster-Shafer formulation allows for consideration of unions of classes, and to represent both imprecision and uncertainty, through the definition of belief and plausibility functions. These two functions, derived from mass function, are generally chosen in a supervised way. In this paper, the authors describe an unsupervised method, based on the comparison of monosource classification results, to select the classes necessary for Dempster-Shafer evidence combination and to define their mass functions. Data fusion is then performed, discarding invalid clusters (e.g. corresponding to conflicting information) thank to an iterative process. Unsupervised multisource classification algorithm is applied to MAC-Europe'91 multisensor airborne campaign data collected over the Orgeval French site. Classification results using different combinations of sensors (TMS and AirSAR) or wavelengths (L- and C-bands) are compared. Performance of data fusion is evaluated in terms of identification of land cover types. The best results are obtained when all three data sets are used. Furthermore, some other combinations of data are tried, and their ability to discriminate between the different land cover types is quantified     

Thermal Microwave Emission from Vegetated Fields: A Comparison between Theory And Experiment

The radiometric measurements over bare field and fields covered with grass, soybean, corn, and alfalfa were made with 1.4-and 5-GHz microwave radiometers during August-October 1978. The measured results are compared with radiative transfer theory treating the vegetated fields as a two-layer random medium. It is found that the presence of a vegetation cover generally gives a higher brightness temperature TB than that expected from a bare soil. The amount of this TB excess increases with increase in the vegetation biomass and in the frequency of the observed radiation. The results of radiative transfer calculations, which include a parameter characterizing ground surface roughness, generally match well with the experimental data.     

基于高光谱影像多维特征的植被精细分类

目前,高光谱植被精细分类存在三个问题:单纯利用光谱信息得到的分类精度较低;光谱数据存在噪声影响了最终的分类结果; 缺少针对具体应用场景而设计的分类方法。为此,提出了一种基于高光谱影像多维特征的植被精细分类方法,通过光谱 数据降维、纹理特征提取以及植被指数选择三个方面对高光谱影像数据进行分析与利用,依靠前期现场调查得到的地面 植被分布情况,选择训练样本并进行支持向量机(Support vector machine, SVM)监督分类,完成地面植被的精细分类, 对分类结果进行验证,总体精度可达99.6\%。结果表明,基于高光谱影像多维特征的植被分类方法能够有效地减小数据噪声、 提高信息利用率,为植被生态监测提供更为准确的数据支撑。     

Assessment of regional biomass-soil relationships using vegetationindexes

The development of photosynthetic active biomass in different ecological conditions, as indicated by normalized difference vegetation indices (NDVIs) is compared by performing a stratified sampling (based on soil associations) on data acquired over Indiana. Data from the NOAA-10 Advanced Very High Resolution Radiometer (AVHRR) were collected for the 1987 and 1988 growing seasons. An NDVI transformation was performed using the two optical bands of the sensor (0.58-0.68 μm and 0.72-1.10 μm). The NDVI is related to the amount of active photosynthetic biomass present on the ground. Statistical analysis of results indicate that land-cover types (forest, forest/pasture, and crops), soil texture, and soil water-holding capacity have an important effect on vegetation biomass changes as measured by AVHRR data     

Effects of Vegetation Cover on the Microwave Radiometric Sensitivity to Soil Moisture

The reduction in sensitivity of the microwave brightness temperature to soil moisture content due to vegetation cover is analyzed using airborne observations made at 1.4 and 5 GHz. The data were acquired during six flights in 1978 over a test site near Colby, Kansas. The test site consisted of bare soil, wheat stubble, and fully mature corn fields. The results for corn indicate that the radiometric sensitivity to soil moisture S decreases in magnitude with increasing frequency and with increasing angle of incidence (relative to nadir).The sensitivity reduction factor, defined in terms of the radiometric sensitivities for bare soil and canopy-covered conditions Y=1 - Scan/ Ss was found to be equal to 0.65 for normal incidence at 1.4 GHz, and increases to 0.89 at 5 GHz. These results confirm previous conclusions that the presence of vegetation cover may pose a serious problem for soil moisture detection with passive microwave sensors.     

Determining land-surface parameters from the ERS wind scatterometer

The ERS-1 wind scatterometer (WSC) has a resolution cell of about 50 km but provides a high repetition rate (less than four days) and makes measurements at multiple incidence angles. In order to retrieve quantitative geophysical parameters over land surfaces using this instrument, a method is presented that applies a mixed-target modeling approach to estimate subpixel fractional vegetation cover at a regional scale. The model represents the footprint area as a combination of part dense, homogeneous vegetation and part bare soil (with homogeneous roughness and dielectric properties). Inversion of this model is then carried out using a retrieval procedure that incorporates a priori information in a quantitative manner The method is applied to the estimation of fractional cover over an area in Africa using WSC data from 1992 to 1995. Retrieved parameters are also compared to ground measurements made in the area during the 1992 HAPEX-Sahel campaign. The procedure illustrates the applicability of WSC data for measuring geophysical parameters over land and offers the potential of deriving a physically-based alternative to empirical indices for estimating regionally-variable parameters     

基于K-means和MTLS-SVM算法的生理参数监测系统

在非医模式的生理参数监测系统中,对监测参数进行学习,可以提高诊断和预测精度.针对多任务时间序列中存在的信息挖掘不充分、预测精度低等问题,将机器学习中的监督和半监督学习方式结合起来对远程健康监护对象进行生理状况预测.该方法用K-means算法将相同类别的数据集群,并使用多任务最小二乘支持向量机(MTLS-SVM)来训练历史数据来进行趋势预测.为了评估该方法的有效性,将MTLS-SVM方法与K-means、MTLS-SVM方法比较,实验结果表明该方法具有较高的预测精度.     

基于可见光光谱图像的红外多光谱图像仿真生成

阐述了红外多光谱图像仿真技术的意义和原理,研究了一种红外多光谱图像的仿真生成方法.提出了一种基于可见光/近红外波段多光谱、超光谱图像数据的地面场景建模方法,以及无监督分类方法和有监督分类方法相结合的地物像元分类、匹配、标记的策略,可以高效地解决像元地物自动匹配标记的问题.利用RGB彩色图像验证了这一方法,在将图像分割后为每类像元赋予相应的红外发射率数值,生成了4个红外波段的多光谱仿真图像,验证了该方法的可行性,指出了多光谱、超光谱图像数据在仿真应用中的各自特点.从仿真结果可以看出:不同波段图像中目标和背景之间呈现不同的特征.该方法可以生成空间形貌和辐射特征接近真实环境的红外多光谱仿真图像,对长波红外波段的多光谱成像探测仪器的研制和目标、背景光谱特征分析与探测算法的研究具有一定意义.     

Effects of Vegetation Cover on the Radar Sensitivity to Soil Moisture

Measurements of the backscattering coefficient ?°, made for bare and vegetation-covered fields, are used in conjunction with a simple backscattering model to evaluate the effects of vegetation cover on the estimation accuracy of soil moisture when derived from radar observations. The results indicate that for soil moisture values below 50 percent of field capacity, the backscatter contribution of the vegetation cover limits the radar's ability to predict soil moisture with an acceptable degree of accuracy. However, for moisture values in the range between 50 and 150 percent of field capacity, the measured ?° is dominated by the soil contribution and the effects of vegetation cover become secondary in importance. It is estimated that in this upper soil moisture range, which is the primary range of interest in hydrology and agriculture, a radar soil moisture prediction algorithm would predict soil moisture with an error of less than ±15 percent of field capacity in 90 percent of the cases.     

Application of neural networks for sea ice classification inpolarimetric SAR images

Several automatic methods have been developed to classify sea ice types from fully polarimetric synthetic aperture radar (SAR) images, and these techniques are generally grouped into supervised and unsupervised approaches. In previous work, supervised methods have been shown to yield higher accuracy than unsupervised techniques, but suffer from the need for human interaction to determine classes and training regions. In contrast, unsupervised methods determine classes automatically, but generally show limited ability to accurately divide terrain into natural classes. In this paper, a new classification technique is applied to determine sea ice types in polarimetric and multifrequency SAR images, utilizing an unsupervised neural network to provide automatic classification, and employing an iterative algorithm to improve the performance. The learning vector quantization (LVQ) is first applied to the unsupervised classification of SAR images, and the results are compared with those of a conventional technique, the migrating means method. Results show that LVQ outperforms the migrating means method, but performance is still poor. An iterative algorithm is then applied where the SAR image is reclassified using the maximum likelihood (ML) classifier. It is shown that this algorithm converges, and significantly improves classification accuracy. The new algorithm successfully identifies first-year and multiyear sea ice regions in the images at three frequencies. The results show that L- and P-band images have similar characteristics, while the C-band image is substantially different. Classification based on single features is also carried out using LVQ and the iterative ML method. It is found that the fully polarimetric classification provides a higher accuracy than those based on a single feature. The significance of multilook classification is demonstrated by comparing the results obtained using four-look and single-look classifications