Decision tree regression for soft classification of remote sensing data

In recent years, decision tree classifiers have been successfully used for land cover classification from remote sensing data. Their implementation as a per-pixel based classifier to produce hard or crisp classification has been reported in the literature. Remote sensing images, particularly at coarse spatial resolutions, are contaminated with mixed pixels that contain more than one class on the ground. The per-pixel approach may result in erroneous classification of images dominated by mixed pixels. Therefore, soft classification approaches that decompose the pixel into its class constituents in the form of class proportions have been advocated. In this paper, we employ a decision tree regression approach to determine class proportions within a pixel so as to produce soft classification from remote sensing data. Classification accuracy achieved by decision tree regression is compared with those achieved by the most widely used maximum likelihood classifier, implemented in the soft mode, and a supervised version of the fuzzy c-means classifier. Root Mean Square Error (RMSE) and fuzzy error matrix based measures have been used for accuracy assessment of soft classification.     

Land cover post-classifications by Markov chain geostatistical cosimulation based on pre-classifications by different conventional classifiers

The recently proposed Bayesian Markov chain random field (MCRF) cosimulation approach, as a new non-linear geostatistical cosimulation method, for land cover classification improvement (i.e. post-classification) may significantly increase classification accuracy by taking advantage of expert-interpreted data and pre-classified image data. The objective of this study is to explore the performance of the MCRF post-classification method based on pre-classification results from different conventional classifiers on a complex landscape. Five conventional classifiers, including maximum likelihood (ML), neural network (NN), Support Vector Machine (SVM), minimum distance (MD), and k-means (KM), were used to conduct land cover pre-classifications of a remotely sensed image with a 90,000 ha area and complex landscape. A sample dataset (0.32% of total pixels) was first interpreted based on expert knowledge from the image and other related data sources, and then MCRF cosimulations were performed conditionally on the expert-interpreted sample dataset and the five pre-classified image datasets, respectively. Finally, MCRF post-classification maps were compared with corresponding pre-classification maps. Results showed that the MCRF method achieved obvious accuracy improvements (ranging from 4.6% to 16.8%) in post-classifications compared to the pre-classification results from different pre-classifiers. This study indicates that the MCRF post-classification method is capable of improving land cover classification accuracy over different conventional classifiers by making use of multiple data sources (expert-interpreted data and pre-classified data) and spatial correlation information, even if the study area is relatively large and has a complex landscape.     

An adaptive spatially constrained fuzzy c-means algorithm for multispectral remotely sensed imagery clustering

This paper presents a novel adaptive spatially constrained fuzzy c-means (ASCFCM) algorithm for multispectral remotely sensed imagery clustering by incorporating accurate local spatial and grey-level information. In this algorithm, a novel weighted factor is introduced considering spatial distance and membership differences between the centred pixel and its neighbours simultaneously. This factor can adaptively estimate the accurate spatial constrains from neighbouring pixels. To further enhance its robustness to noise and outliers, a novel prior probability function is developed by integrating the mutual dependency information in the neighbourhood to obtain accurate spatial contextual information. The proposed algorithm is free of any experimentally adjusted parameters and totally adaptive to the local image content. Not only the neighbourhood but also the centred pixel terms of the objective function are all accurately estimated. Thus, the ASCFCM enhances the conventional fuzzy c-means (FCM) algorithm by producing homogeneous regions and reducing the edge blurring artefact simultaneously. Experimental results using a series of synthetic and real-world images show that the proposed ASCFCM outperforms the competing methodologies, and hence provides an effective unsupervised method for multispectral remotely sensed imagery clustering.     

Estimating impervious surfaces from medium spatial resolution imagery: a comparison between fuzzy classification and LSMA

Impervious surfaces are important environmental indicators and are related to many environmental issues, such as water quality, stream health and the urban heat island effect. Therefore, detailed impervious surface information is crucial for urban planning and environment management. To extract impervious surfaces from remote sensing imagery, many algorithms and techniques have been developed. However, there are still debates over the strengths and limitations of linear versus nonlinear algorithms in handling mixed pixels in the urban landscapes. In the meantime, although many previous studies have compared various techniques, few comparisons were made between linear and nonlinear techniques. The objective of this study is to compare the performance between nonlinear and linear methods for impervious surface extraction from medium spatial resolution imagery. A linear spectral mixture analysis (LSMA) and a fuzzy classifier were applied to three Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) images acquired on 5 April 2004, 16 June 2001 and 3 October 2000, which covered Marion County, Indiana, United States. An aerial photo of Marion County with a spatial resolution of 0.14 m was used for validation of estimation results. Six impervious surface maps were yielded, and an accuracy assessment was performed. The root mean square error (RMSE), the mean average error (MAE), and the coefficient of determination (R ²) were calculated to indicate the accuracy of impervious surface maps. The results show that the fuzzy classification outperformed LSMA in impervious surface estimation in all seasons. For the June image, LSMA yielded a result with an RMSE of 13.2%, while the fuzzy classifier yielded an RMSE of 12.4%. For the April image, LSMA yielded an accuracy of 21.1% and the fuzzy classifier yielded 17.0%. For the October image, LSMA yielded a result with an RMSE of 19.8%, but the fuzzy classifier yielded an RMSE of 17.5%. Moreover, a subset image of the commercial, high-density and low-density residential areas was selected in order to compare the effectiveness of the developed algorithms for estimating impervious surfaces of different land use types. The result shows that the fuzzy classification was more effective than LSMA in both high-density and low-density residential areas. These areas prevailed with mixed pixels in the medium resolution imagery, such as ASTER. The results from the tested commercial area had a very high RMSE value due to the prevalence of shade in the area. It is suggested that the fuzzy classifier based on the nonlinear assumption can handle mixed pixels more effectively than LSMA.     

基于熵加权K-means全局信息聚类的高光谱图像分类

目的 高光谱图像波段数目巨大，导致在解译及分类过程中出现“维数灾难”的现象。针对该问题，在K-means聚类算法基础上，考虑各个波段对不同聚类的重要程度，同时顾及类间信息，提出一种基于熵加权K-means全局信息聚类的高光谱图像分类算法。方法 首先，引入波段权重，用来刻画各个波段对不同聚类的重要程度，并定义熵信息测度表达该权重。其次，为避免局部最优聚类，引入类间距离测度实现全局最优聚类。最后，将上述两类测度引入K-means聚类目标函数，通过最小化目标函数得到最优分类结果。结果 为了验证提出的高光谱图像分类方法的有效性，对Salinas高光谱图像和Pavia University高光谱图像标准图中的地物类别根据其光谱反射率差异程度进行合并，将合并后的标准图作为新的标准分类图。分别采用本文算法和传统K-means算法对Salinas高光谱图像和Pavia University高光谱图像进行实验，并定性、定量地评价和分析了实验结果。对于图像中合并后的地物类别，光谱反射率差异程度大，从视觉上看，本文算法较传统K-means算法有更好的分类结果；从分类精度看，本文算法的总精度分别为92.20%和82.96%， K-means算法的总精度分别为83.39%和67.06%，较K-means算法增长8.81%和15.9%。结论 提出一种基于熵加权K-means全局信息聚类的高光谱图像分类算法，实验结果表明，本文算法对高光谱图像中具有不同光谱反射率差异程度的各类地物目标均能取得很好的分类结果。     

Approaches for the production and evaluation of fuzzy land cover classifications from remotely-sensed data

Remote sensing is an attractive source of data for land cover mapping applications. Mapping is generally achieved through the application of a conventional statistical classification, which allocates each image pixel to a land cover class. Such approaches are inappropriate for mixed pixels, which contain two or more land cover classes, and a fuzzy classification approach is required. When pixels may have multiple and partial class membership measures of the strength of class membership may be output and, if strongly related to the land cover composition, mapped to represent such fuzzy land cover. This type of representation can be derived by softening the output of a conventional ‘hard’ classification or using a fuzzy classification. The accuracy of the representation provided by a fuzzy classification is, however, difficult to evaluate. Conventional measures of classification accuracy cannot be used as they are appropriate only for ‘hard’ classifications. The accuracy of a classification may, however, be indicated by the way in which the strength of class membership is partitioned between the classes and how closely this represents the partitioning of class membership on the ground. In this paper two measures of the closeness of the land cover representation derived from a classification to that on the ground were used to evaluate a set of fuzzy classifications. The latter were based on measures of the strength of class membership output from classifications by a discriminant analysis, artificial neural network and fuzzy c-means classifiers. The results show the importance of recognising and accommodating for the fuzziness of the land cover on the ground. The accuracy assessment methods used were applicable to pure and mixed pixels and enabled the identification of the most accurate land cover representation derived. The results showed that the fuzzy representations were more accurate than the ‘hard’ classifications. Moreover, the outputs derived from the artificial neural network and the fuzzy c-means algorithm in particular were strongly related to the land cover on the ground and provided the most accurate land cover representations. The ability to appropriately represent fuzzy land cover and evaluate the accuracy of the representation should facilitate the use of remote sensing as a source of land cover data.     

Mapping sub-pixel proportional land cover with AVHRR imagery

A problem with NOAA AVHRR imagery is that the intrinsic scale of spatial variation in land cover in the U.K. is usually finer than the scale of sampling imposed by the image pixels. The result is that most NOAA AVHRR pixels contain a mixture of land cover types (sub-pixel mixing). Three techniques for mapping the sub-pixel proportions of land cover classes in the New Forest, U.K. were compared: (i) artificial neural networks (ANN); (ii) mixture modelling; and (iii) fuzzy c -means classification. NOAA AVHRR imagery and SPOT HRV imagery, both for 28 June 1994, were obtained. The SPOT HRV images were classified using the maximum likelihood method, and used to derive the 'known' sub-pixel proportions of each land cover class for each NOAA AVHRR pixel. These data were then used to evaluate the predictions made (using the three techniques and the NOAA AVHRR imagery) in terms of the amount of information provided, the accuracy with which that information is provided, and the ease of implementation. The ANN was the most accurate technique, but its successful implementation depended on accurate co-registration and the availability of a training data set. Supervised fuzzy c -means classification was slightly more accurate than mixture modelling.     

Segmentation of remotely-sensed images by a split-and-merge process+

Abstract

This paper describes the application of an image segmentation technique to remotely-sensed terrain images used for environmental monitoring. The segmentation is a preprocessing operation which is applied prior to image classification in order to improve classification accuracy from that achievable by classifying pixels individually on the basis of their spectral signatures. The method uses a split-and-merge technique to segment images into regions of homogeneous tone and texture wherever this is possible. The split-and-merge technique employs a hierarchical quadtree data structure. Texture is measured using easily computed grey value difference statistics. The homogeneity criteria employed in region merging are dependent on local statistics. The segmented image is classified using a region classifier for regions and the normal per-pixel classifier for single pixels in areas of inhomogeneity. The technique is illustrated by example classifications of aerial Multispectral Scanner data from two test sites. A quantitative analysis of the performance shows that an increased classification accuracy is achieved.     

Automatic image pixel clustering with an improved differential evolution

This article proposes an evolutionary-fuzzy clustering algorithm for automatically grouping the pixels of an image into different homogeneous regions. The algorithm does not require a prior knowledge of the number of clusters. The fuzzy clustering task in the intensity space of an image is formulated as an optimization problem. An improved variant of the differential evolution (DE) algorithm has been used to determine the number of naturally occurring clusters in the image as well as to refine the cluster centers. We report extensive performance comparison among the new method, a recently developed genetic-fuzzy clustering technique and the classical fuzzy c-means algorithm over a test suite comprising ordinary grayscale images and remote sensing satellite images. Such comparisons reveal, in a statistically meaningful way, the superiority of the proposed technique in terms of speed, accuracy and robustness.     

Performance characteristics of the indicator classifier on simulated image data

Classification of remotely sensed imagery into groups of pixels having similar spectral reflectance characteristics is conducted classically by comparing the spectral signature of unknown pixels with those of training pixels of known ground cover type. Thus classification methods use only the spectral characteristics of image data without considering the spatial aspects or the relative location of an unknown pixel with respect to pixels from the training data set. An indicator classifier was introduced in 1992 that combines spatial and spectral information in a decision model. In this Letter the performance of this classifier is tested on simulated image data with known mineral targets and varying spatial variability and noise. It is demonstrated that incorporating spatial continuity into the classification process may largely increase the accuracy of the resulting classified images.     

An automatic method for mapping inland surface waterbodies with Radarsat-2 imagery

The use of synthetic aperture radar (SAR) imagery is generally considered to be an effective method for detecting surface water. Among various supervised/unsupervised classification methods, a SAR-intensity-based histogram thresholding method is widely used to distinguish waterbodies from land. A SAR texture-based automatic thresholding method is presented in this article. The use of texture images substantially enhances the contrast between water and land in intensity images. It also makes the method less sensitive to incidence angles than intensity-based methods. A modified Otsu thresholding algorithm is applied to selected sub-images to determine the optimal threshold value. The sub-images were selected using k-means results to ensure a sufficient number of pixels for both water and land classes. This is critical for the Otsu algorithm being able to detect an optimal threshold for a SAR image. The method is completely unsupervised and is suitable for large SAR image scenes. Tests of this method on a Radasat-2 image mosaicked from 8 QuadPol scenes covering the Spritiwood valley in Manitoba, Canada, show a substantial increase in land–water classification accuracy over the commonly used SAR intensity thresholding method (kappa indices are 0.89 vs. 0.79). The method is less computationally intensive and requires less user interaction. It is therefore well suited for detecting waterbodies and monitoring their dynamic changes from a large SAR image scene in a near-real time environment).     

An automatic region-based image segmentation algorithm for remote sensing applications

Object-based image analysis has proven its potentials for remote sensing applications, especially when using high-spatial resolution data. One of the first steps of object-based image analysis is to generate homogeneous regions from a pixel-based image, which is typically called the image segmentation process. This paper introduces a new automatic Region-based Image Segmentation Algorithm based on k-means clustering (RISA), specifically designed for remote sensing applications. The algorithm includes five steps: k-means clustering, segment initialization, seed generation, region growing, and region merging. RISA was evaluated using a case study focusing on land-cover classification for two sites: an agricultural area in the Republic of South Africa and a residential area in Fresno, CA. High spatial resolution SPOT 5 and QuickBird satellite imagery were used in the case study. RISA generated highly homogeneous regions based on visual inspection. The land-cover classification using the RISA-derived image segments resulted in higher accuracy than the classifications using the image segments derived from the Definiens software (eCognition) and original image pixels in combination with a minimum-distance classifier. Quantitative segmentation quality assessment using two object metrics showed RISA-derived segments successfully represented the reference objects.     

An effective hybrid approach to remote-sensing image classification

This article presents a hybrid fuzzy classifier for effective land-use/land-cover (LULC) mapping. It discusses a Bayesian method of incorporating spatial contextual information into the fuzzy noise classifier (FNC). The FNC was chosen as it detects noise using spectral information more efficiently than its fuzzy counterparts. The spatial information at the level of the second-order pixel neighbourhood was modelled using Markov random fields (MRFs). Spatial contextual information was added to the MRF using different adaptive interaction functions. These help to avoid over-smoothing at the class boundaries. The hybrid classifier was applied to advanced wide-field sensor (AWiFS) and linear imaging self-scanning sensor-III (LISS-III) images from a rural area in India. Validation was done with a LISS-IV image from the same area. The highest increase in accuracy among the adaptive functions was 4.1% and 2.1% for AWiFS and LISS-III images, respectively. The paper concludes that incorporation of spatial contextual information into the fuzzy noise classifier helps in achieving a more realistic and accurate classification of satellite images.     

基于蜂群[k]-means算法的遥感图像聚类应用研究

在遥感领域，获取用于训练的标记数据耗费巨大且困难，因此许多非监督技术逐渐被发展和应用于标记样本有限的遥感图像。将[k]均值和蜂群算法相结合，提出一种新的非监督聚类算法。使用灰度共生矩阵和小波变换提取遥感图像特征，对特征数据集进行蜂群[k]-means聚类。整个聚类过程首先使用最大最小距离积邻域均值法产生初始聚类中心，将蜂群算法和[k]-means算法交替执行，实现遥感图像的聚类。通过UCI数据集和凉水国家级自然保护区的遥感数据的实验结果表明，该算法具有较高的聚类准确率，满足遥感图像聚类的应用需求。     

An Improved Method for Edge Detection and Image Segmentation Using Fuzzy Cellular Automata

Image segmentation is one of the most important and challenging problems in image processing. The main purpose of image segmentation is to partition an image into a set of disjoint regions with uniform attributes. In this study, we propose an improved method for edge detection and image segmentation using fuzzy cellular automata. In the first stage, we introduce a new edge detection method based on fuzzy cellular automata, called the texture histogram, and empirically demonstrate the efficiency of the proposed method and its robustness in denoising images. In the second stage, we propose an edge detection algorithm by considering the mean values of the edges matrix. In this algorithm, we use four fuzzy rules instead of 32 fuzzy rules reported earlier in the literature. In the third and final stage, we use the local edge in the edge detection stage to more accurately accomplish image segmentation. We demonstrate that the proposed method produces better output images in comparison with the separate segmentation and edge detection methods studied in the literature. In addition, we show that the method proposed in this study is more flexible and efficient when noise is added to an image.     

基于模糊高斯基函数神经网络的遥感图像分类

针对遥感图像分类的特点,提出了一种基于模糊高斯基函数神经网络的遥感图像分类器。该分类器将模糊技术与神经网络相结合,采用神经网络来实现模糊推理,利用神经网络的学习能力来达到调整模糊隶属函数和模型规则的目的,从而使系统具备了自适应的特性,实验结果表明,这种基于模糊高斯基孙数神经网络的分类器经过训练后,可应用于遥感图像的分类,其分类精度明显高于传统的最大似然分类法。     

Texture classification and segmentation using wavelet packet frame and Gaussian mixture model

In this paper, we propose a scheme for texture classification and segmentation. The methodology involves an extraction of texture features using the wavelet packet frame decomposition. This is followed by a Gaussian-mixture-based classifier which assigns each pixel to the class. Each subnet of the classifier is modeled by a Gaussian mixture model and each texture image is assigned to the class to which pixels of the image most belong. This scheme shows high recognition accuracy in the classification of Brodatz texture images. It can also be expanded to an unsupervised texture segmentation using a Kullback-Leibler divergence between two Gaussian mixtures. The proposed method was successfully applied to Brodatz mosaic image segmentation and fabric defect detection.     

Contextual classification of Landsat TM images to forest inventory cover types

In the context of Landsat TM images forest stands are a cluster of homogeneous pixels. Contextual classification of forest cover types exploits relationships between neighbouring pixels in the pursuit of an increase in classification accuracy. Results with six contextual classifiers from two sites in Canada were compared to results with a maximum likelihood (ML) classifier. The comparisons were done at three levels of spectral class separation. Training and validation data were obtained from single-stage cluster sampling of 2?km×2?km primary sampling units (PSU) located on a 20?km×20?km grid. A strong relationship between contextual and ML classification accuracy was explored with logistic regression analysis. Effects of contextual classification were predicted for given levels of ML accuracy. Estimates of the spatial autocorrelation of reflectance values within a PSU were deemed consistent with a first-order autoregressive process. Iterative Conditional Modes (ICM) was the best contextual method; it improved the overall accuracy by four to six percentage points (statistically significant) when ML accuracy was between 50% and 80%. A relaxed ICM and a smoothing algorithm were second and third best. Contextual classification is most promising when an ML accuracy is around 70%. ICM results were sensitive to the level of spatial autocorrelation of ML classification errors and to the homogeneity of a PSU.     

自然场景图像与合成图像的快速分类

目的 随着现代通信和传感技术的快速发展,互联网上多媒体数据日益增长,既为人们生活提供了便利,又给信息有效利用提出了挑战。为充分挖掘网络图像中蕴含的丰富信息,同时考虑到网络中图像类型的多样性,以及不同类型的图像需要不同的处理方法,本文针对当今互联网中两种主要的图像类型：自然场景图像与合成图像,设计层次化的快速分类算法。方法 该算法包括两层,第1层利用两类图像在颜色,饱和度以及边缘对比度上表现出来的差异性提取全局特征,并结合支持向量机（SVM）进行初步分类,第1层分类结果中低置信度的图像会被送到第2层中。在第2层中,系统基于词袋模型（bag-of-words）对图像不同类型的局部区域的纹理信息进行编码得到局部特征并结合第2个SVM分类器完成最终分类。针对层次化分类框架,文中还提出两种策略对两个分类器进行融合,分别为分类器结果融合与全局+局部特征融合。为测试算法的实用性,同时收集并发布了一个包含超过30 000幅图像的数据库。结果 本文设计的全局与局部特征对两类图像具有较强的判别性。在单核Intel Xeon（R）（2.50 GHz）CPU上,分类精度可达到98.26%,分类速度超过40帧/s。另外通过与基于卷积神经网络的方法进行对比实验可发现,本文提出的算法在性能上与浅层网络相当,但消耗更少的计算资源。结论 本文基于自然场景图像与合成图像在颜色、饱和度、边缘对比度以及局部纹理上的差异,设计并提取快速有效的全局与局部特征,并结合层次化的分类框架,完成对两类图像的快速分类任务,该算法兼顾分类精度与分类速度,可应用于对实时性要求较高的图像检索与数据信息挖掘等实际项目中。