Spatial-Aware Supervised Learning for Hyper-Spectral Image Classification Comprehensive Assessment

A comprehensive assessment of the spatial.aware mpervised learning algorithms for hyper.spectral image （HSI） classification was presented. For this purpose, standard support vector machines （ SVMs ）, mudttnomial logistic regression （ MLR ） and sparse representation （SR） based supervised learning algorithm were compared both theoretically and experimentally. Performance of the discussed techniques was evaluated in terms of overall accuracy, average accuracy, kappa statistic coefficients, and sparsity of the solutions. Execution time, the computational burden, and the capability of the methods were investigated by using probabilistie analysis. For validating the accuracy a classical benchmark AVIRIS Indian pines data set was used. Experiments show that integrating spectral.spatial context can further improve the accuracy, reduce the misclassltication error although the cost of computational time will be increased.

High-resolution Hyper-spectral Image Classification with Parts-based Feature and Morphology Profile in Urban Area

High-resolution hyper-spectral image (HHR) provides both detailed structural and spectral information for urban study. However, due to the inherent correlation between spectral bands and within-class variability in the data, the data processing of HHR is a challenging work. In this paper, based on spectral mixture analysis theory, a new stack of parts description features were extracted, and then incorporated with a stack of morphology based spatial features. Partially supervised constrained energy minimization (CEM) and unsupervised nonnegative matrix factorization (NMF) were used to extract the part-features. The joint features were then integrated by SVM classifier. The advantages of this method are the representation of physical composition of the urban area by the parts-features and the show of multi-scale structure information by morphology profiles. Experiments with an airborne hyper-spectral data flightline over the Washington DC Mall were performed, and the performance of the proposed algorithm was evaluated in comparison with well-known nonparametric weighted feature extraction (NWFE) and feature selection method. The results shown that the proposed features-joint scheme consistently outperforms the traditional methods, and so can provide an effective option for processing HHR data in urban area.

Distance-based separability criterion of ROI in classification of farmland hyper-spectral images

The hyper-spectral image contains spectral and spatial information,which increases the ability and precision of objects classification.Despite the classification value of hyper-spectral imaging technology within various applications,users often find it difficult to effectively apply in practice because of the effect of light,temperature and wind in outdoor environment.This research presented a new classification model for outdoor farmland objects based on near-infrared(NIR)hyper-spectral images.It involves two steps including region of interest(ROI)acquisition and establishment of classifiers.A distance-based method for quantitative analysis was proposed to optimize the reference pixels in ROI acquisition firstly.Then maximum likelihood(ML)and support vector machine(SVM)were used for farmland objects classification.The performance of the proposed method showed that the total classification accuracy based on the reference pixels was over 97.5%,of which the SVM-M model could reach 99.5%.The research provided an effective method for outdoor farmland image classification.

Segmentation and classification of high resolution imagery for mapping individual species in a closed canopy,deciduous forest

In this paper we investigate the use of a shadow-based delineation program for identifying segments in imagery of a closed canopy, deciduous forest, in West Virginia, USA, as a way to reduce the noise associated with per-pixel classification in forested environments. Shadows typically cluster along the boundaries of trees and therefore can be used to provide a network of nodes for the delineation of segments. A minimum cost path algorithm, where cost is defined as the cumulative sum of brightness values traversed along the connecting route, was used to connect shadow clumps. To test this approach, a series of classifications was undertaken using a multispectral digital aerial image of a six hectare test site and a minimum cost path segmentation. Three species were mapped: oaks, red maple and yellow poplar. The accuracy of an aspatial maximum likelihood classification (termed PERPIXEL classification) was 68.5%, compared to 74.0% for classification using the mean vector of the segments identified with the minimum cost path algorithm (MEAN_SEG), and 78% when the most common class present in the segment is assigned to the entire segment (POSTCLASS_SEG). By comparison, multispectral classification of the multispectral data using the field-mapped polygons of individual trees as segments, produced an accuracy of 82.3% when the mean vector of the polygon was used for classification (MEAN_TREE), and 85.7% when the most common class was assigned to the entire polygon (POSTCLASS_TREE). A moving window-based post-classification majority filter (POSTCLASS_MAJ5BY5) produced an intermediate accuracy value, 73.8%. The minimum cost path segmentation algorithm was found to correctly delineate approximately 28% of the trees. The remaining trees were either segmented, aggregated, or a combination of both segmented and aggregated. Varying the threshold that was used to discriminate shadows appeared to have little effect on the number of correctly delineated trees, or on the overall accuracy of the multispectral classification, although it did have a notable effect on the proportions of aggregated and Segmented trees.

基于模糊C均值聚类的林地遥感图像分类研究

在以往国内外相关研究的基础上,以我国东北长白山系典型林区为试验区,以2007年7月Landat5卫星TM多光谱图像为遥感数据,运用模糊C均值聚类方法对遥感图像进行分类试验.分类结果显示:模糊C均值分类方法在总分类精度和Kappa系数上都占有一定的优势,比传统分类方法有着更好的分类效果.模糊C均值方法在林地植被的遥感分类中具有较好的应用前景.

基于词袋模型的林业业务图像分类

针对林业业务图像的特点,提出了一种基于稠密尺度不变特征转换(Dense SIFT)特征的词袋(BoW)模型,并联合直方图正交核的支持向量机(SVM)对图像自动分类。首先采用Dense SIFT提取林业业务图像特征,然后使用BoW模型描述各业务图像,最后利用SVM进行分类识别。实验结果表明:采用Dense SIFT特征比SIFT特征训练时间和识别时间更短,并有更高的识别率,更适应实时性较高的场合;SVM采用多项式核函数(Poly),径向基核函数(RBF),多层感知器核函数(Sigmoid)以及直方图交叉核对3类林业业务图像分类时,直方图正交核取得的平均识别率最高;综合Dense SIFT在局部特征上的优势,加上BoW模型和直方图交叉核SVM分类器,平均识别率达到了86.7%,有较好的识别效果。

基于分形维度的林业遥感图像树种分类识别

传统的树种分类识别方法未进行最大池化操作,导致树种分类识别精度差。现引入分形维度进行林业遥感图像树种分类识别。通过ROI区域截取获取遥感树种图像,利用直方图均衡化方法进行原始图像预处理,以便获得高质量与清晰度的林业遥感图像;通过分形维度理论分析提取的林业遥感图像纹理特征,完成卷积神经网络模型的优化构建;将林业遥感图像纹理特征输入卷积层,经卷积层的卷积操作并计算特征数据,池化池通过最大池化操作卷积层输出的数据;通过Relu激活函数对林业遥感图像树种纹理特征进行深度分析,利用Softmax分类器实现树种分类识别。实验结果表明,上述方法预处理后的遥感图像质量高,且林业遥感图像树种分类识别的效率高,分类识别的时间低至35.7ms,分类识别的准确率高达95.62%。

基于Multi-CNN空间特征提取的高光谱遥感影像分类

在遥感影像处理领域,对高光谱遥感影像分类处理的需求日益增长,由于大量的高光谱遥感影像训练样本获得较难,使得卷积神经网络不适合应用到高光谱遥感影像分类中。针对此问题,文中提出一种基于Multi-CNN空间特征提取的高光谱遥感影像分类模型,该模型将原始高光谱遥感影像作为输入,最终的分类结果作为输出;自动从不同的尺度提取输入数据的空间特征;解决获得大量有标记高光谱遥感影像训练样本的棘手问题;通过伊春凉水林场数据集上的实验结果表明,文中建立的分类模型,在分类正确率上优于其他分类模型,分类正确率达到92.31%。

基于多源遥感数据的数字林场信息提取与系统构建

数字林业是林业现代化的基础。本研究以河北省塞罕坝机械林场为研究区,利用遥感科学与技术,基于机载数据(激光雷达、高分辨率CCD影像和高光谱影像),研究单木分割和单木分类方法,实现了机载飞行区域270 km2范围内所有单木的信息(位置、树高、冠幅、树种等)提取。结合地面样方调查数据,单木分割匹配精度可以达到0.6以上,研究区内4种典型树种单木分类精度均能达到97%以上。基于地基激光雷达扫描数据,研究以单木为单位,通过枝干重建和叶片添加,实现单木三维模型重建和场景重建。该方法有利于从器官(枝干、叶片)—单木—区域多尺度分析森林结构特征。在此基础上,本研究探索了以单木为基础的林业数字化方法,并结合WebGIS技术开发了数字林业系统,初步实现了在研究区内单棵树为单位的数据存储、管理、查询、分析和可视化等功能,为后续的森林精准化经营管理和决策规划提供重要的基础数据和系统支持。