Journal of Geo-information Science >
The Theory and Calculation of Spatial-spectral Cognition of Remote Sensing
Received date: 2016-01-04
Request revised date: 2016-03-11
Online published: 2016-05-10
Copyright
In recent years, with the rapid development of earth observation technologies, remote sensing using the satellites has gradually entered the era of big data. Facing the current demands and characteristics of remote sensing applications, it is feasible and necessary to explore the theories and methods of high-spatial-resolution remote sensing cognition with the cooperation of visual cognition. In this context, we are inspired by Geo-informatic-Tupu and intend to study the spatial-spectral cognition of remote sensing. This paper systematically presents the theory and calculation methodology for the spatial-spectral cognition of remote sensing, and expects to standardize the processes of remote sensing information extraction, as a result to further build a sophisticated, quantitative, intelligent and integrated model for the remote sensing information interpretation. The whole methodology contains two directions' cognitive calculation, namely horizontal "bottom-up hierarchical abstraction" and longitudinal "top-down knowledge transfer". These two steps are corresponded with three principal Spatial-Spectral transformation processes, which are summarized as "extracting spatial maps based on clustering pixels' spectrum", "coordinating spatial-spectral features" and "understanding attributes through the recognition of known diagram". Our study focuses on the analysis of the involved concepts, the basic idea, the key technologies and their existing difficulties, and emphasizes on the utilization of big data and gradually the application of integrated knowledge to achieve different levels of remote sensing cognition. Through these approaches, we expect to provide a new perspective for the remote sensing interpretation with the adoption of big data resources.
LUO Jiancheng , WU Tianjun , XIA Liegang . The Theory and Calculation of Spatial-spectral Cognition of Remote Sensing[J]. Journal of Geo-information Science, 2016 , 18(5) : 578 -589 . DOI: 10.3724/SP.J.1047.2016.00578
Fig. 1 Spatial-spectral explanation of remote sensing information图1 遥感信息的图谱解释 |
Fig. 2 The flowchart of spatial-spectral transition in the spatial-spectral cognition of remote sensing图2 遥感图谱认知中的图谱转化 |
Fig. 3 Extracting spatial maps based on clustering pixels’ spectrum图3 由谱聚图 |
Tab. 1 The key technologies and their existing difficulties they have in the stage of “extracting spatial maps based on clustering pixels’ spectrum”表1 由谱聚图阶段的关键技术及其难点问题 |
关键技术 | 常用实现算法 | 存在问题与难点 | 发展趋势 |
---|---|---|---|
分割 | 基于边缘、区域(阈值、图论、能量泛函)的多种分割算法 | 一般的分割算法对地物复杂多变的遥感影像适用性较低,且数据量巨大的高分影像使分割效率大幅下降,如何提升分割的效率 | 发展复杂环境下高分辨率影像的多尺度快速聚合技术(多尺度:大数据综合处理) |
均值漂移、分水岭等多尺度分割算法 | 如何设置合适的尺度集来合理地表达地物的异尺度特征,实现成功抽取对象的目标[44],即如何提升分割的精度 | ||
聚类、分类 | Kmeans、ISODATA等非监督分类算法的聚类,基于SVM、神经网络、决策树、随机森林等监督分类算法的像元级分类或对象级分类 | 像元级分类造成的椒盐噪声的影响;对象级分类受分割算法问题的影响,存在对象分离不合理的问题;对象合并规则的设定 | |
人工矢量编辑 | 目视勾画 | 矢量编辑工具的智能化程度,减少人工操作量 |
Fig. 4 Coordinating spatial-spectral features图4 图谱协同 |
Tab. 2 The key technologies and their existing difficulties in the stage of “coordinating spatial-spectral features”表2 图谱协同阶段的关键技术及其难点问题 |
关键技术 | 常用实现算法 | 存在问题与难点 | 发展趋势 |
---|---|---|---|
特征分析 | 光谱、空间、时间、地域等多源“图-谱”特征的提取与优选算法 | 地物特征的计算具有不确定性,造成后续的分类存在一定的错分率 | ① 构建“影像-结构-演化”紧致结构的多特征表达模型(多特征:异构时空特征表达) ② 建立高空间、高光谱与高时间分辨率大数据协同计算(多维度:多源数据协同处理) |
多时相获取的遥感数据是非平稳信号,地物特征具有不一致性,当协同时序特征时,需对数据进行有有效、可靠的滤波去噪、几何配准、辐射校正等一致性预处理 | |||
分类 | SVM、神经网络、决策树、随机森林等监督分类算法以及自训练算法、生成模型、图论方法、多视角算法等半监督分类算法 | 地物外在特征不能描述其本质特征,典型特征难以确定,限制了分类的精细度(即地物可分性和分类精确性)[45] | |
训练样本的采集是费时费力的步骤,如何在少量样本或无样本条件下实现高精度的分类、提升分类的自动化与智能化程度[46] |
Fig. 5 Understanding attributes through the recognition of known diagram图5 认图知谱 |
Tab. 3 The key technologies and their existing difficulties in the stage of “understanding attributes through the recognition of known diagram”表3 认图知谱阶段的关键技术及其难点问题 |
关键技术 | 常用实现算法 | 存在问题与难点 | 发展趋势 |
---|---|---|---|
传统地物识别技术 | 分割、特征提取、分类等算法 | 同表1、表2的说明 | ① 建立波谱、视觉、环境和空间知识的逐步融合模型(多知识:遥感与GIS一体化) ② 发展“谱相”及“图形”间螺旋式认知的自适应计算模型(多模型:流程化、自动化) |
迁移学习 | 实例迁移、特征迁移、参数迁移、关系知识迁移等算法 | 知识的形式化以及如何在不同时间、空间、尺度的先验知识中进行去伪存真以及与任务的关联 | |
GIS空间分析 | 空间关系分析、叠置分析、网络分析、缓冲分析、地统计分析等算法 | GIS数据与所需提取遥感信息的关联分析,以及各种GIS空间分析方法中存在的限制问题 | |
语义推理 | 特征编码或表达(视觉词袋)、主题模型(概率潜语义分析pLSA、潜在狄利克雷分析LDA)、深度特征学习等算法 | 底层特征与高层语义间的鸿沟[8] |
Fig.6 Syllogism structure in the spatial-spectral cognition of remote sensing图6 遥感图谱认知的三段论结构 |
Fig. 7 Calculations in the spatial-spectral cognition of remote sensing图7 遥感图谱认知计算 |
Fig.8 Hierarchical abstraction in the spatial-spectral cognition of remote sensing图8 遥感认知计算中的分层抽象 |
Tab. 4 The form, calculation and expression of knowledge in the spatial-spectral cognition of remote sensing表4 遥感图谱认知过程中知识的形态、计算和表达 |
来源 | 形态 | 表达 | 层次 | 计算 | 存储 | 迁移 | 应用 | |
---|---|---|---|---|---|---|---|---|
遥感知识 | 视觉知识 | 影像(对象)特征:色调、几何形状、纹理等波谱和和空间形态特征等 | 低 | 图像处理统计 | 对象表达特征库(表) | 特征迁移 | 由谱聚图 | |
参数知识 | 影像参数:成像时间、角度、传感器性能等 | 低 | 查询查阅 | 参数文档 | 参数迁移 | |||
地域知识 | 波谱知识 | 地物波谱库、纯端元地物样本库等 | 中 | 采集测量 | 波谱库(表) | 特征迁移 | 由图聚图、 图谱协同 | |
环境知识 | 坡度、坡向等DEM高程相关信息;温度、湿度等土地资源信息; | 中 | 采集测量 经验总结 | 带有环境特征的栅格/矢量图斑、If… Then…规则 | 特征/关系知识迁移 | |||
解译知识 | 模型知识 | 物理量反演模型、专题指数计算模型、定理分析模型、分类体系等 | 高 | 实验分析 模型解算 | 公式、文档 | 参数迁移 | 图谱协同 | |
物候知识 | 作物生长演变地学规律(季相变化物候特征)、景观格局演变规律等 | 高 | 时序分析 | 特征变化曲线 | 参数迁移 | |||
空间知识 | 空间分布 | 地类的地理空间分布:土地利用/覆盖解译图的图斑、空间样本位置库 | 高 | 目视/机器解译 | 带有空间位置和地类属性的栅格/矢量图斑 | 关系知识迁移 | ||
空间关系 | 阴影等相邻、相交、包含、方向等空间关系/格局知识 | 高 | GIS 空间分析 | 语义网络、拓扑 | 关系知识迁移 |
The authors have declared that no competing interests exist.
[1] |
[
|
[2] |
|
[3] |
|
[4] |
[
|
[5] |
[
|
[6] |
|
[7] |
|
[8] |
|
[9] |
|
[10] |
|
[11] |
|
[12] |
|
[13] |
|
[14] |
|
[15] |
|
[16] |
|
[17] |
|
[18] |
|
[19] |
|
[20] |
|
[21] |
|
[22] |
|
[23] |
|
[24] |
|
[25] |
|
[26] |
|
[27] |
|
[28] |
|
[29] |
|
[30] |
[
|
[31] |
[
|
[32] |
|
[33] |
|
[34] |
|
[35] |
|
[36] |
[
|
[37] |
[
|
[38] |
[
|
[39] |
[
|
[40] |
[
|
[41] |
[
|
[42] |
[
|
[43] |
[
|
[44] |
[
|
[45] |
[
|
[46] |
[
|
/
〈 |
|
〉 |