遥感影像居民地信息自动提取方法研究进展

快速掌握居民地的时空分布和变化特征对于区域可持续发展具有重要意义。遥感技术的发展，为居民地的快速提取提供了新的途径。传统的人工目视解译费时费力，基于遥感影像的居民地信息自动提取势在必行。主要探讨了居民地信息自动提取的5种方法，即基于统计的分类法、谱间结构阀值法、归一化指数法、纹理分析法、地学专家知识分类法，讨论了各种方法的优缺点，总结了目前的研究进展。最后提出了居民地信息自动提取方法的发展趋势．认为多种分类方法相结合，建立居民地信息提取的专家系统将是今后研究的重点。     

Expert Classification Method Based on Patch-Based Neighborhood Searching Algorithm

The expert knowledge has been widely used to improve the remotely sensed classification accuracy. Generally, the ex-pert classification system mainly depends on DEM and some thematic maps. The spatial relationship information in pixel level was commonly introduced into the expert classification. Because the geographic objects were found spatially dependent relationship to a certain degree, the commonly used basic unit of spatial relationship information in pixel greatly limited the efficiency of spatial in-formation. A patch-based neighborhood searching algorithm was proposed to implement the expert classification. The homogene-ous spectral unit, patch, was used as the basic unit in the spatial object granularity, and different types of patches' relationship in-formation were obtained through a spatial neighborhood searching algorithm. And then the neighborhood information and DEM data were added into the expert classification system and used to modify the primitive classification errors. In this case, the classi-fication accuracies of wetland, grassland and cropland were obviously improved. In this work, water was used as base object, and different types of water extraction methods were tested to get a result in a high accuracy.     

抚州地区TM影像的水土流失信息自动提取专家模式研究

从遥感信息分析入手，分析了抚州地区水土流失区在Landsat TM1,TM2,TM3,TM4,TM5,TM6,TM7各个波段上与其它背景地物的可分性。利用水土流失区光谱特征与其它地物的差异性，建立条件表达式，并建立简单模型，即可把该区的水土流失区提取出来。该区土地利用率不是很高，水土流失比较严重，而且面积大。由于居民地光谱特征与水土流失区比较接近，所以区分居民地与水土流失区成为研究的重点。研究表明，通过建立简单的模型，可以将水土流失区提取出来，并能达到很好的效果。     

赣东林地专家分类方法初探

遥感技术在解决森林调查方面有其独到的优势。通过对赣东地区林地和农作物分布地区的光谱特征的分析研究，建立了适合于该地区的林地专家知识分类,模型自动提取和统计了各类林地的面积与森林覆盖率。与人工调查相比，该方法精度高且省时省力。其精度可达70％以上。     

廊固凹陷大兴砾岩体隐蔽油气藏储层预测研究

廊固凹陷大兴断层下降盘沙三下砾岩体具有储层埋藏深、非均质性强的隐蔽性油气藏特征，近年来钻探成功率较低。针对该区地质条件，需充分借助综合地质专家系统，采取分级约束沆程开展砾岩体储层地震预测研究，可揭示砾岩体储层的空间展布和油气分布规律，明确大兴砾岩体下步滚动方向，并有效地降低了勘探风险。     

遥感图像理解专家系统中面向对象的知识表示

阐述了在遥感图像理解专家系统中，用面向对象的方法表示知识，使得图像分类效率更高     

The FAM(Fuzzy Asociative Memory)neural network model and its application in earthquake prediction

ＴｈｅＦＡＭ（ＦｕｚｚｙＡｓｏｃｉａｔｉｖｅＭｅｍｏｒｙ）ｎｅｕｒａｌｎｅｔｗｏｒｋｍｏｄｅｌａｎｄｉｔｓａｐｐｌｉｃａｔｉｏｎｉｎｅａｒｔｈｑｕａｋｅｐｒｅｄｉｃｔｉｏｎＷＥＩＷＡＮＧ１）（王炜）ＧＥＮＧＦＥＮＧＷＵ２）（吴耿锋）ＢＩＮＳＨＵＨ...     

Influence of neighbourhood information on ‘Local Climate Zone’ mapping in heterogeneous cities

Local climate zone (LCZ) mapping is an emerging field in urban climate research. LCZs potentially provide an objective framework to assess urban form and function worldwide. The scheme is currently being used to globally map LCZs as a part of the World Urban Database and Access Portal Tools (WUDAPT) initiative. So far, most of the LCZ maps lack proper quantitative assessment, challenging the generic character of the WUDAPT workflow. Using the standard method introduced by the WUDAPT community difficulties arose concerning the built zones due to high levels of heterogeneity. To overcome this problem a contextual classifier is adopted in the mapping process. This paper quantitatively assesses the influence of neighbourhood information on the LCZ mapping result of three cities in Belgium: Antwerp, Brussels and Ghent. Overall accuracies for the maps were respectively 85.7 ± 0.5, 79.6 ± 0.9, 90.2 ± 0.4%. The approach presented here results in overall accuracies of 93.6 ± 0.2, 92.6 ± 0.3 and 95.6 ± 0.3% for Antwerp, Brussels and Ghent. The results thus indicate a positive influence of neighbourhood information for all study areas with an increase in overall accuracies of 7.9, 13.0 and 5.4%. This paper reaches two main conclusions. Firstly, evidence was introduced on the relevance of a quantitative accuracy assessment in LCZ mapping, showing that the accuracies reported in previous papers are not easily achieved. Secondly, the method presented in this paper proves to be highly effective in Belgian cities, and given its open character shows promise for application in other heterogeneous cities worldwide.     

Using spectrum differentiation and combination for target detection of minerals

Among the techniques that have been developed in spectroscopy, derivative analysis is particularly promising for use with remote sensing data. In the first step of this research we apply the derivative spectrum in a real hyperspectral image and introduce a new target detection approach called “DCEM”. For this purpose, 1st to 5th orders of derivative spectrum were applied to the DCEM. The outcome of this research has shown that the application of derivative spectrum in target detection is perfectly advisable in a specific derivative order for each target. This order can be introduced as an optimized order or the Best DCEM. The spectrum differentiation eliminates low frequency components of the spectrum. Despite the little information included in those low frequency components of a signal or spectrum, their complete elimination cause an information loss problem. Hence, in the second step of this research an ensemble classifier approach was employed for the combined use of both spectra and the best derivative order. This simultaneous use of the derivative and zero order spectra is introduced as “ECEM”. Experiments were conducted via a HyMap hyperspectral airborne image in eastern Iran. The detection results show that both proposed methods significantly outperform CEM in ROC and AUC values. The best performance upgrade in DCEM detection was about 24% for Kaolinite target and about 28% for Alunite target in ECEM.     

深度学习的半监督遥感图像检索

遥感图像数据的海量性、多样性和复杂性等特点对遥感图像检索的速度和精度提出了更高的要求,其中特征提取是影响遥感图像检索效果的关键。本文方法首先对遥感图像进行预处理,然后基于稀疏自动编码的方法在大量未标注的遥感图像上进行特征学习得到特征字典,基于卷积神经网络的思想,使用训练出来的特征字典对遥感图像进行卷积和池化得到每幅图像的特征图;接下来使用特征图训练Softmax分类器;最后对待检索图像分类,在同一类别中计算特征间的距离,进而实现遥感图像的检索。实验结果表明,该方法能够有效提高遥感图像检索的速度和准确度。