Entropy-based fuzzy classification parameter optimization using uncertainty variation across spatial resolution

In the past researchers have suggested hard classification approaches for pure pixel remote sensing data and to handle mixed pixels soft classification approaches have been studied for land cover mapping. In this research work, while selecting fuzzy c-means (FCM) as a base soft classifier entropy parameter has been added. For this research work Resourcesat-1 (IRS-P6) datasets from AWIFS, LISSIII and LISS-IV sensors of same date have been used. AWIFS and LISS-III datasets have been used for classification and LISS-III and LISS-IV data were used for reference data generation, respectively. Soft classified outputs from entropy based FCM classifiers for AWIFS and LISS-III datasets have been evaluated using sub-pixel confusion uncertainty matrix (SCM). It has been observed that output from FCM classifier has higher classification accuracy with higher uncertainty but entropy-based classifier with optimum value of regularizing parameter generates classified output with minimum uncertainty.     

ALCM: Automatic land cover mapping

It may be quite important for resource management people to extract single land cover class, at sub-pixel level from multi-spectral remote sensing images of different areas in single step processing. It has been observed, that neural network can be trained to extract single land cover class from multi-spectral remote sensing images, but they have problems in setting various parameters and slow during training stage. This paper present single land cover class water, extraction from mixed pixels present in multiple multi-spectral remote sensing data sets of same bands of AWiFS sensor of Resoursesat-1 (IRS-P6) satellite from different areas. In this work fuzzy logic-based algorithm, which is independent of statistical distribution assumption of data, has been studied at sub-pixel level to handle mixed pixels. It has been found; possibilistic c-means (PCM) algorithm takes the possibilistic view, that the membership of a feature vector in a class has nothing to do with its membership in other classes. Due to this, it was observed that PCM can extract only one class, from remote sensing multi-spectral data and it has produced 93.7% and 97.1% overall sub-pixel classification accuracy for two different data sets of different places using LISS-III (IRS-P6) reference data of same dates as of AWiFS data.     

Sub-pixel land-cover mapping with improved fraction images upon multiple-point simulation

Outputs of soft classification inherently contain uncertainty. As an input for the sub-pixel mapping (SPM) method, the uncertainty is propagated to SPM result especially the boundary region between classes. Therefore, reducing the uncertainty within the outputs of soft classification is worth exploring. This paper firstly utilizes multiple-point simulation (MPS) through training images for characterizing the spatial structural properties of a surface object/class. Consequently, MPS results are used to increase the accuracy of the fraction image of the surface object/class. The improved fraction image then inputs to the SPM method for producing the land cover map with finer spatial resolution. In order to validate the proposed method, a remotely sensed image from Landsat TM 30 m over the Qianyanzhou red earth hill region in China is used. This experimental study not only compares the results from SPM with improved fraction images with MPS and results from SPM with original fraction images, but also investigates the performances of different soft classifiers. It has been demonstrated that this proposed method is an effective way to reduce the uncertainty in outputs of different soft classification, increase the recognition accuracies of boundary regions and thus increase the accuracies of SPM simulated images.     

Adaptive MAP sub-pixel mapping model based on regularization curve for multiple shifted hyperspectral imagery

Sub-pixel mapping is a promising technique for producing a spatial distribution map of different categories at the sub-pixel scale by using the fractional abundance image as the input. The traditional sub-pixel mapping algorithms based on single images often have uncertainty due to insufficient constraint of the sub-pixel land-cover patterns within the low-resolution pixels. To improve the sub-pixel mapping accuracy, sub-pixel mapping algorithms based on auxiliary datasets, e.g., multiple shifted images, have been designed, and the maximum a posteriori (MAP) model has been successfully applied to solve the ill-posed sub-pixel mapping problem. However, the regularization parameter is difficult to set properly. In this paper, to avoid a manually defined regularization parameter, and to utilize the complementary information, a novel adaptive MAP sub-pixel mapping model based on regularization curve, namely AMMSSM, is proposed for hyperspectral remote sensing imagery. In AMMSSM, a regularization curve which includes an L-curve or U-curve method is utilized to adaptively select the regularization parameter. In addition, to take the influence of the sub-pixel spatial information into account, three class determination strategies based on a spatial attraction model, a class determination strategy, and a winner-takes-all method are utilized to obtain the final sub-pixel mapping result. The proposed method was applied to three synthetic images and one real hyperspectral image. The experimental results confirm that the AMMSSM algorithm is an effective option for sub-pixel mapping, compared with the traditional sub-pixel mapping method based on a single image and the latest sub-pixel mapping methods based on multiple shifted images.     

Soft classification for sub-pixel land cover extraction

With increasing resolution of the remotely sensed data the problems of images contaminated by mixed pixels arc frequent. Conventional classification techniques often produce erroneous results when applied to images dominated by mixed pixels. This may load to unrealistic representation of land cover, thereby, affecting efficient planning, management and monitoring of natural resources. Consequently, soft classification techniques providing sub-pixel land cover information may have to be utilised. From a range of soft classification techniques, the present study focuses on the utility of conventional maximum likelihood classifier and linear mixture modelling for sub-pixel. land cover classifications. The accuracy of the soft classifications has been assessed using distance measures and correlation co-efficient. The results show that linear mixture modelling has produced accuracies comparable to maximum likelihood classifier. Besides this the correlations between actual land cover proportions and proportions from linear mixture modelling, though not strong, arc statistically significant at 95% level of confidence. It has also been observed that the normalised likelihoods of maximum likelihood classifier also show strong correlations with the actual land cover proportions on ground and therefore has the potential to be used as a soft classification technique.     

The effect of imposing ‘fractional abundance constraints’ onto the multilayer perceptron for sub-pixel land cover classification

To be physically interpretable, sub-pixel land cover fractions or abundances should fulfill two constraints, the Abundance Non-negativity Constraint (ANC) and the Abundance Sum-to-one Constraint (ASC). This paper focuses on the effect of imposing these constraints onto the MultiLayer Perceptron (MLP) for a multi-class sub-pixel land cover classification of a time series of low resolution MODIS-images covering the northern part of Belgium. Two constraining modes were compared, (i) an in-training approach that uses ‘softmax’ as the transfer function in the MLP’s output layer and (ii) a post-training approach that linearly rescales the outputs of the unconstrained MLP. Our results demonstrate that the pixel-level prediction accuracy is markedly increased by the explicit enforcement, both in-training and post-training, of the ANC and the ASC. For aggregations of pixels (municipalities), the constrained perceptrons perform at least as well as their unconstrained counterparts. Although the difference in performance between the in-training and post-training approach is small, we recommend the former for integrating the fractional abundance constraints into MLPs meant for sub-pixel land cover estimation, regardless of the targeted level of spatial aggregation.     

元胞自动机的遥感影像混合像元分类

通过对元胞自动机理论的研究,提出元胞自动机的遥感影像混合像元分解模型。利用多波段遥感数据验证混合像元分解算法的可行性,并将结果与线性分解模型进行比较。结果表明,元胞自动机混合像元分解模型在分解的准确性方面,明显优于一般线性模型的精度。最后,将分类结果与传统的监督分类算法比较,得出元胞自动机的混合像元分解模型明显优于监督分类精度的结论。     

Distribution of Errors in a Classified Map of Satellite Data

Abstract

The output from any spatial data processing method may contain some uncertainty. With the increasing use of satellite data products as a source of data for Geographical Information Systems (GIS), there have been some major concerns about the accuracy of the satellite‐based information. Due to the nature of spatial data and remotely sensed data acquisition technology, and conventional classification, any single classified image can contain a number of mis‐classified pixels. Conventional accuracy evaluation procedures can report only the number of pixels that are mis‐classified based on some sampling observation. This study investigates the spatial distribution and the amount of these pixels associated with each cover type in a product of satellite data. The study uses Thematic Mapper (TM) and SPOT multispectral data sets obtained for a study area selected in North East New South Wales, Australia. The Fuzzy c‐Means algorithm is used to identify the classified pixels that contained some uncertainty. The approach is based on evaluating the strength of class membership of pixels. This study is important as it can give an indication of the amount of error resulting from the mis‐classification of pixels of specific cover types as well as the spatial distribution of such pixels. The results show that the spatial distribution of erroneously classified pixels are not random and varies depending on the nature of cover types. The proportions of such pixels are higher in spectrally less clearly defined cover types such as grasslands.     

基于SVM的高光谱遥感图像亚像元定位

提出了基于支持向量机（support vector machine,SVM）的高光谱遥感图像亚像元定位方法。全变分（total variation,TV）模型是经典的保边缘平滑滤波器,本文将其引入作为预处理,来提高混合像元分解及亚像元定位的精度;本文方法在训练和检验样本的构建过程中,依据空间相关性理论,同时考虑了中心像元及其邻近像元丰度值对亚像元类别归属的影响;在监督分类训练和检验过程中,通过剔除纯净像元来缩减样本数量,在保证算法准确性的同时提高了效率。对真实高光谱遥感数据进行了实验,主观评价和定量分析验证了本文方法的有效性。     

Reservoir capacity estimation of Singoor reservoir,India, by per-pixel and sub-pixel processing of multi-date satellite images

Traditional approaches of image classification, such as maximum likelihood and the band thresholding method, involve the per-pixel approach to delineate the water spread area of a reservoir. One of the limitations of these approaches is that the pixels representing the reservoir border, containing a mixture of water, soil and vegetation, are classified entirely as water, thereby resulting in inaccurate estimates of the water spread area. To compute the water spread area accurately, the sub-pixel approach has been used in this study. The water spread areas extracted using per-pixel and sub-pixel approaches from IRS-1D and P6 satellite image data were in turn used to quantify the capacity of the Singoor reservoir, Andhra Pradesh, India. The estimated capacity of the reservoir using the per-pixel and sub-pixel approaches was 727.75 Mm³ and 716.11 Mm³, respectively. The validation shows that the sub-pixel approach produced much less error (1.08%) than the per-pixel based approach (3.14%).     

面向高光谱图像分类的半监督空谱判别分析

为充分利用高光谱图像蕴藏的空间信息提升分类精度,提出了面向高光谱图像分类的半监督空谱判别分析(S3 DA)算法。考虑高光谱图像数据集的空间一致性,首先利用少量标记样本定义类内散度矩阵,保存数据集同类像元的光谱近邻结构;再利用无标记样本定义空间近邻像元散度矩阵,揭示像元间的空间近邻结构和地物的空间分布结构信息。S3 DA既保持数据集在光谱域的可分性,又保存了无标记样本蕴藏的空间域近邻结构,增强了同类像元和空间近邻像元在投影子空间的聚集性,从而提升分类性能。在PaviaU和Indian Pines数据集的试验表明,总体分类精度分别达到81.50%和71.77%。与传统的光谱方法比较,该算法能有效提升高光谱图像数据集的地物分类精度。     

Estimation of Water-Spread Area of Singoor Reservoir,Southern India by Super Resolution Mapping of Multispectral Satellite Images

Estimation of reservoir water-spread area is often carried out by field surveys which are cumbersome, time consuming, expensive and involves more man power. Hence, such surveys cannot be carried out periodically. To overcome this issue, satellite images are used, wherein the reservoir water-spread is estimated by conventional per-pixel classification such as the maximum likelihood and minimum distance to mean approaches that often results in inaccurate estimate of water-spread area due to the presence of mixed pixels. High cost and non-availability of high resolution images demands the use of an alternative approach that can give accurate information about the reservoir water-spread area. In this work, IRS, LISS III images of moderate (24 m) resolution were used for accurate estimation of the water-spread area of Singoor reservoir, southern India. The reservoir water-spread areas were extracted using per-pixel classification, sub-pixel classification and super resolution mapping approaches. These results were validated with the water-spread areas obtained from field data of the same dates. The error produced by the per-pixel approach was 6.66%, while it was 4.37% for the sub-pixel approach and a meagre 1.71% for the super-resolution approach. Fairly less error produced by the super resolution mapping technique implies that it is an efficient approach for accurate quantification of reservoir water-spread area. The estimated water-spread can be used in a simple volume estimation formula to estimate the capacity of the reservoir.     

结合纹理的SVM遥感影像分类研究

针对传统统计模式识别分类方法分类精度不高,分类时未加入像元灰度的空间分布和结构特征以及分类时样本不足等缺陷,采用一种结合纹理的支持向量机(SVM)遥感图像分类方法。该方法在对Landsat7 ETM遥感影像进行纹理特征提取的基础上,构建了结合纹理的SVM分类模型。以河南省汝阳县为试验区,利用此模型对该区域的土地利用类型进行分类研究,并将分类结果与最大似然法和单源数据(光谱)SVM分类结果进行定性和定量比较分析。研究结果表明:该方法能够有效地解决单数据源分类效果破碎、分类精度不高等问题;对高维输入向量具有较高的推广能力;总精度达到90%,比单源信息的SVM分类法提高了6%,而与最大似然法相比,总精度提高了近9%,取得了良好的效果。     

Evaluation of fuzzy-based classifiers for cotton crop identification

In this study, an evaluation of fuzzy-based classifiers for specific crop identification using multi-spectral temporal data spanning over one growing season has been carried out. The temporal data sets have been georeferenced with 0.3 pixel rms error. Temporal information of cotton crop has been incorporated through the following five indices: simple ratio (SR), normalized difference vegetation index (NDVI), transformed normalized difference vegetation index (TNDVI), soil-adjusted vegetation index (SAVI) and triangular vegetation index (TVI), to study the effect of indices on classified output. For this purpose, a comparative study between two fuzzy-based soft classification approaches, possibilistic c-means (PCM) and noise classifier (NC), was undertaken. In this study, advanced wide field sensor (AWiFS) data for soft classification and linear imaging self scanner sensor (LISS III) data for soft testing purpose from Resourcesat-1 (IRS-P6) satellite were used. It has been observed that NC fuzzy classifier using TNDVI temporal index – dataset 2, which comprises four temporal images performs better than PCM classifier giving highest fuzzy overall accuracy of 96.03%.     

专家系统在 TM 图像分类中的应用

提高TM图像的分类精度,是图像处理及应用领域中一个很重要的研究课题。本文在总结已有成果基础上,首先利用现有的统计分类技术,对待分类图像进行预分类,并检测出“不确定”像元。然后综合光谱、地理、土壤类型、早期判别结果、目视判读经验等各种知识和信息,充分发挥专家系统的推理判断能力,对“不确定”像元的类别作进一步判别,使得整幅图像的分类精度得到改善。并据此初步建立了一个土地利用的分类系统。试验证明,这种分类方法的精度比仅用单一多光谱信息的统计分类法(最大似然法)提高约8％。     

组合核支持向量回归提取高光谱影像不透水面

由于城市地表组成的复杂性,基于单核函数的支持向量回归模型很难满足精度。本文结合空间-光谱组合核函数和支持向量回归,提出了一种提取高光谱影像不透水面丰度的改进算法。首先从高光谱遥感图像上提取波谱特征和多通道灰度共生矩阵空间纹理特征,选取研究区10%像元特征数据作为训练数据,以线性加权求和核为多核组合方式,建立结合光谱信息和空间信息的组合核支持向量回归模型。然后,用生成的回归模型预测未知像元不透水面丰度值。最后,对实验结果进行评价。在模拟数据试验中,本文算法比单核回归均方根误差平均降低1.4%,决定系数比单核回归平均提高0.6%。在Hyperion数据两组试验中,该算法比单核回归均方根误差平均降低1.8%,决定系数比单核回归平均提高11.7%。模拟和真实两种高光谱数据实验中,本文算法均得到了空间形态上更准确的不透水面结果,单核回归结果存在失真现象。研究结果表明:本文算法能够有效提取城市不透水面丰度,与单核方法相比有较明显的精度提升。     

Temporal MODIS data for identification of wheat crop using noise clustering soft classification approach

In this study, temporal MODIS-Terra MOD13Q1 data have been used for identification of wheat crop uniquely, using the noise clustering (NC) soft classification approach. This research also optimises the selection of date combination and vegetation index for classification of wheat crop. First, a separability analysis is used to optimise the date combination for each case of number of dates and vegetation index. Then, these scenes have undergone for NC soft classification. The resolution parameter (δ) was optimised for the NC classifier and found to be a value of 1.6 × 10⁴ for wheat crop identification. Classified outputs were analysed by receiver operating characteristics (ROC) analysis for sub-pixel detection. Highest area under the ROC curve was found for soil-adjusted vegetation index corresponding to the three different phenological stages data sets. From this study, the data sets corresponding to the Sowing, Flowering and Maturity phenological stages of wheat crop were found more suitable to identify it uniquely.     

Multi-resolution integration of land cover for sub-pixel estimation of urban impervious surface and forest cover

Abstract

A methodology is presented for estimating percent coverage of impervious surface (IS) and forest cover (FC) within Landsat thematic mapper (TM) pixels of urban areas. High-resolution multi-spectral images from Quickbird (QB) play a key role in the sub-pixel mapping process by providing information on the spatial distributions of ISs and FCs at 2.4 m ground sampling intervals. Thematic classifications, also derived from the Landsat imagery, have then been employed to define relationships between 30 m Landsat-derived greenness values and percent IS and FC. By also utilizing land cover/land use classification derived from Landsat and defining unique relationships for urban sub-classes (i.e. residential, commercial/industrial, open land), confusion between impervious and fallow agricultural lands has been overcome. Test results are presented for Ottawa-Gatineau, an urban area that encompasses many aspects typical of the North American urban landscape. Multiple QB scenes have been acquired for this urban centre, thereby allowing us to undertake an in-depth study of the error budgets associated with the fractional inference process.     

The use of MODIS data to derive acreage estimations for larger fields: A case study in the south-western Rostov region of Russia

Recent developments in remote sensing technology, in particular improved spatial and temporal resolution, open new possibilities for estimating crop acreage over larger areas. Remotely sensed data allow in some cases the estimation of crop acreage statistics independently of sub-national survey statistics, which are sometimes biased and incomplete. This work focuses on the use of MODIS data acquired in 2001/2002 over the Rostov Oblast in Russia, by the Azov Sea. The region is characterised by large agricultural fields of around 75 ha on average. This paper presents a methodology to estimate crop acreage using the MODIS 16-day composite NDVI product. Particular emphasis is placed on a good quality crop mask and a good quality validation dataset. In order to have a second dataset which can be used for cross-checking the MODIS classification a Landsat ETM time series for four different dates in the season of 2002 was acquired and classified. We attempted to distinguish five different crop types and achieved satisfactory and good results for winter crops. Three hundred and sixty fields were identified to be suitable for the training and validation of the MODIS classification using a maximum likelihood classification. A novel method based on a pure pixel field sampling is introduced. This novel method is compared with the traditional hard classification of mixed pixels and was found to be superior.