基于多源数据源的森林资源年度动态监测研究
|
摘要
森林资源监测是林业重要的基础性工作,是国情国力调查的重要组成部分。传统森林资源监测的间隔周期一般为5—10年,信息产出的时效性相对滞后,但是随着遥感RS技术、GIS技术和GPS技术在森林资源监测中的综合研究深入和不断进展,基于遥感技术的森林资源的动态监测正在受到世界各国越来越多的重视。本研究旨在提高森林资源动态监测能力,增强森林资源监督管理的针对性和时效性,选取辽宁省鞍山市为示范研究区,采用多源、多级分辨率、多时相的遥感数据,结合地面调查数据,通过遥感反演技术来获得森林相关参数(例如：森林面积、林分郁闭度、森林蓄积量等),利用实地调查数据对模型计算得到的森林参数进行检验,通过多时相的遥感数据,对所获得的森林参数进行动态分析,完成对示范监测区森林资源动态变化的监测工作。研究结果表明：
(1)在现行森林资源调查方法的基础上,以PPS抽样为基础,综合利用多级遥感数据,提出基于抽样技术的多级遥感监测技术方法,并以鞍山市为例,利用2006年TM、MODIS和航片等遥感数据,进行多级遥感监测技术试验,试验结果表明,与传统遥感方法相比,可以大幅提高MODIS低分辨率影像提取大区域森林分布信息的精度,分类精度达到86.77%。
(2)以TM为主要数据源,利用线性光谱解混模型对研究区三期影像的郁闭度进行估测,预测值与野外实测值比较正确率达到80%以上,生成研究区三期郁闭度图层,并进行了森林郁闭度动态变化分析,结果显示：在1995至2006年期间,鞍山市林分郁闭度呈现先下降后上升的趋势,但林分整体郁闭度水平1995年为最好。
(3)采用分水岭算法,利用航片影像资料,通过一类调查样地法和目视样地法进行单木森林结构参数提取实验,提取了单木的株树、树冠的平均面积,树冠的平均冠幅,以及单木的位置等信息,并对实验结果进行了分析和讨论。结果显示,利用高分辨影像资料,采用分水岭算法,可以有效提取单木森林结构参数(株数、位置、冠幅等),这种方法可应用于少林地区或城市区域的森林资源调查。
(4)以三期TM为主要数据源,结合一类调查样地数据,采用逐步回归方法进行森林蓄积量遥感估测模型的建模实验。利用所建立的遥感估测模型,分别对鞍山市1995年、2000年、2006年3期森林蓄积量进行了估测,并分析了其时空动态变化特
征。结果表明,三期模型得到的森林蓄积量预测值分别为688.11×105m3、755.86×
105m3和1042.67×105m3；说明鞍山市森林蓄积量从1995至2006年呈现增长的趋势；
并利用2007年森林资源二类调查小班统计蓄积进行验证,2006年模型估测值与2007
小班蓄积统计值相比,两者相差13.79%,模型精度达到85%以上,较为理想。(5)采用一类调查数据资料,利用立地分级方法建立平均树高、平均胸径、公顷
株数和公顷蓄积的生长模型,通过所建立的生长模型,可应用于森林资源调查中相关
因子的更新。同时,通过研究森林蓄积、林分生长量和消耗量与年龄、平均树高、平
均胸径和郁闭度等等林分因子之间的关系,建立了森林蓄积、生长量和消耗量的估测
模型。结果表明,基于生长模型来更新森林资源档案数据,实现森林资源的年度监测
的方法是可行的。(6)研究发展了基于多源数据源的森林资源遥感技术方法,建立了市级森林资源
年度动态遥感监测体系,并成功应用于鞍山市的森林资源年度动态监测中,取得了很
好的效果。通过鞍山市示范应用,能及时掌握鞍山市2007-2009年森林资源动态变化,
为鞍山市制定林业发展、生态环境宏观规划提供可靠的依据。本研究首次综合利用多源、多级遥感数据,结合地面调查样地资料,采用基于
PPS抽样技术的多级遥感监测方法,对研究区森林面积进行定量提取研究。在森林资
源动态监测中,采用遥感技术和空间信息技术为采集数据的主要手段,遥感宏观监测
与典型地面调查相结合,定量分析与定性评价相结合,模型分析与实地测量相结合的
方法,解决了目前森林资源监测迫切需要的关键技术,建立了鞍山市森林资源年度监
测技术体系,并成功应用于鞍山市森林资源的年度动态监测中,大大提高了鞍山市森
林资源监测能力,增强森林资源监督管理的针对性和时效性,降低野外调查费用,满
足市级森林资源日常调查业务、行政性监管与执法职能的要求,填补市级森林资源年
度监测技术的空白。
Forest resource monitoring is an important basic work of forestry, also an important part of our national situation. Due to the long monitoring period for5to10years and lagging information output of traditional forest resource, it is hard to meet the needs of the forestry development and ecological construction. With the deeper research and application of RS, GIS and GPS technology in forest resources monitoring, dynamic monitoring based on remote sensing technology is brought to the attention of the countries all over the world more and more. This study aims to improve the dynamic monitoring ability of forest resource, which helps strengthen the pertinence and efficiency of forest resources supervision and management. We selected Anshan (in Liaoning province) as demonstration research area, used multi-source, multi-resolution, multi-temporal remote sensing data, combined with field inventory data, to obtain some indexes (such as forest area, canopy density, volume, etc.) through remote sensing inversion technique. Then, the field survey data were used for model testing. Meanwhile, the dynamic of forest resource indexes were analysised by multi-temporal remote sensing data. Finally, we completed the monitoring work of dynamic changes for forest resources at the demonstration zone of Anshan. The results showed that:
(1) Based on the current investigation method for forest resources and-PPS sampling, with comprehensive utilization of multiple remote sensing data, multi-level remote sensing method based on sampling technology was put forward. Taking Anshan as an example, TM, MODIS and aerial photographs of2006were used for the multi-level remote sensing experiment. Results showed that, compared with the traditional remote sensing method, the new method can significantly improve the classification accuracy of large area forest distribution by low resolution MODIS images for, which reached to86.77%.
(2) Taking TM image as main source of data, and using linear spectral mixing model, the crown density indexes in the images at3stages for the studied area were evaluated. Comparing with field measured values, the prediction accuracy was up to80%. Crown density layers of3stages were generated then, for the analysis on the dynamic changes of forest canopy density. The results indicated that forest crown density in Aanshan showed a trend of "U-shaped" during1995to2006, with a best level in1995.
(3) Based on the data of aerial photography, with field inventory and visual plot materials, forest structural indexes were extracted by watershed algorithm. The extraction of tree number, canopy coverage, crown width, and location for single tree were analyzed and discussed then. Results showed the forest structural indexes can be extracted effectively, and it could be applied in forest survey for open forestland or urban area.
(4) Taking TM images at3stages in1995,2000and2006,as main data source, combined with the national forest inventory data, forest stock volume was estimated by stepwise regression method. The space-time dynamic progress of stock estimations for3stages were analysed respectively. Results showed a growth trend for forest stock volume from1995to2006, with estimations followed by688.11×105m3、755.86×105m3and1042.67×105m3. Comparing estimated volume in2006with sub-comparment statistic result of that in2007from forest resource inventory planning and design, we found no significant difference between them (13.79%), which suggested the model can ideally estimate volume, with high accuracy (above85%).
(5) Based on the national forest inventory data, growth models for height, DBH (diameter at breast height) and volume were finished by site classification method. All of that growth models were used for the updates of tree measured factors in forest resources survey. Meanwhile, prediction models for volume, growth and consumption at stand level were bulited with some related factors as age, height, DBH crown density, etc. Results showed that archival data of forest resources can be updated based on growth model, which would helped for the annual monitoring of forest resources.
(6). Our research improved the technology of forest resources monitoring based on Multi-data Source, and built a city level system for the annual monitoring of forest resources, which had been applicated successfully in Anshan. With the demonstration application, we learned the dynamic of forest resource from2007to2009, and also provided reliable basis and construction suggestions for forestry development and eco-environment planning.
(7) Our study promoted to comprehensive utilization of multi-source, multi-level remote sensing data. We studied quantitative extraction of forest area by multilevel remote sensing method based on PPS sampling, and combined with field survey data. Focused on dynamic monitoring of forest resource, we took remote sensing technology and spatial information technology as main methods for gathering data. And then, a system for the annual monitoring of forest resources in Anshan was built by series methods as remote sensing macroscopic monitoring&field investigation, quantitative analysis&qualitative analysis, model estimation&field measurements. This study improved the urgently need technology to monitor forest resource dynamic, and successfully applied to Anshan annual monitoring of forest resources dynamic. In conclusion, our research filled the blank of the municipal annual monitoring technology of forest resources. The techonology could greatly improve the ability of forest resources monitoring in Anshan, reduce field investigation cost, and met the need of municipal survey, supervision and enforcement of forest resources.
(1)在现行森林资源调查方法的基础上,以PPS抽样为基础,综合利用多级遥感数据,提出基于抽样技术的多级遥感监测技术方法,并以鞍山市为例,利用2006年TM、MODIS和航片等遥感数据,进行多级遥感监测技术试验,试验结果表明,与传统遥感方法相比,可以大幅提高MODIS低分辨率影像提取大区域森林分布信息的精度,分类精度达到86.77%。
(2)以TM为主要数据源,利用线性光谱解混模型对研究区三期影像的郁闭度进行估测,预测值与野外实测值比较正确率达到80%以上,生成研究区三期郁闭度图层,并进行了森林郁闭度动态变化分析,结果显示：在1995至2006年期间,鞍山市林分郁闭度呈现先下降后上升的趋势,但林分整体郁闭度水平1995年为最好。
(3)采用分水岭算法,利用航片影像资料,通过一类调查样地法和目视样地法进行单木森林结构参数提取实验,提取了单木的株树、树冠的平均面积,树冠的平均冠幅,以及单木的位置等信息,并对实验结果进行了分析和讨论。结果显示,利用高分辨影像资料,采用分水岭算法,可以有效提取单木森林结构参数(株数、位置、冠幅等),这种方法可应用于少林地区或城市区域的森林资源调查。
(4)以三期TM为主要数据源,结合一类调查样地数据,采用逐步回归方法进行森林蓄积量遥感估测模型的建模实验。利用所建立的遥感估测模型,分别对鞍山市1995年、2000年、2006年3期森林蓄积量进行了估测,并分析了其时空动态变化特
征。结果表明,三期模型得到的森林蓄积量预测值分别为688.11×105m3、755.86×
105m3和1042.67×105m3；说明鞍山市森林蓄积量从1995至2006年呈现增长的趋势；
并利用2007年森林资源二类调查小班统计蓄积进行验证,2006年模型估测值与2007
小班蓄积统计值相比,两者相差13.79%,模型精度达到85%以上,较为理想。(5)采用一类调查数据资料,利用立地分级方法建立平均树高、平均胸径、公顷
株数和公顷蓄积的生长模型,通过所建立的生长模型,可应用于森林资源调查中相关
因子的更新。同时,通过研究森林蓄积、林分生长量和消耗量与年龄、平均树高、平
均胸径和郁闭度等等林分因子之间的关系,建立了森林蓄积、生长量和消耗量的估测
模型。结果表明,基于生长模型来更新森林资源档案数据,实现森林资源的年度监测
的方法是可行的。(6)研究发展了基于多源数据源的森林资源遥感技术方法,建立了市级森林资源
年度动态遥感监测体系,并成功应用于鞍山市的森林资源年度动态监测中,取得了很
好的效果。通过鞍山市示范应用,能及时掌握鞍山市2007-2009年森林资源动态变化,
为鞍山市制定林业发展、生态环境宏观规划提供可靠的依据。本研究首次综合利用多源、多级遥感数据,结合地面调查样地资料,采用基于
PPS抽样技术的多级遥感监测方法,对研究区森林面积进行定量提取研究。在森林资
源动态监测中,采用遥感技术和空间信息技术为采集数据的主要手段,遥感宏观监测
与典型地面调查相结合,定量分析与定性评价相结合,模型分析与实地测量相结合的
方法,解决了目前森林资源监测迫切需要的关键技术,建立了鞍山市森林资源年度监
测技术体系,并成功应用于鞍山市森林资源的年度动态监测中,大大提高了鞍山市森
林资源监测能力,增强森林资源监督管理的针对性和时效性,降低野外调查费用,满
足市级森林资源日常调查业务、行政性监管与执法职能的要求,填补市级森林资源年
度监测技术的空白。
Forest resource monitoring is an important basic work of forestry, also an important part of our national situation. Due to the long monitoring period for5to10years and lagging information output of traditional forest resource, it is hard to meet the needs of the forestry development and ecological construction. With the deeper research and application of RS, GIS and GPS technology in forest resources monitoring, dynamic monitoring based on remote sensing technology is brought to the attention of the countries all over the world more and more. This study aims to improve the dynamic monitoring ability of forest resource, which helps strengthen the pertinence and efficiency of forest resources supervision and management. We selected Anshan (in Liaoning province) as demonstration research area, used multi-source, multi-resolution, multi-temporal remote sensing data, combined with field inventory data, to obtain some indexes (such as forest area, canopy density, volume, etc.) through remote sensing inversion technique. Then, the field survey data were used for model testing. Meanwhile, the dynamic of forest resource indexes were analysised by multi-temporal remote sensing data. Finally, we completed the monitoring work of dynamic changes for forest resources at the demonstration zone of Anshan. The results showed that:
(1) Based on the current investigation method for forest resources and-PPS sampling, with comprehensive utilization of multiple remote sensing data, multi-level remote sensing method based on sampling technology was put forward. Taking Anshan as an example, TM, MODIS and aerial photographs of2006were used for the multi-level remote sensing experiment. Results showed that, compared with the traditional remote sensing method, the new method can significantly improve the classification accuracy of large area forest distribution by low resolution MODIS images for, which reached to86.77%.
(2) Taking TM image as main source of data, and using linear spectral mixing model, the crown density indexes in the images at3stages for the studied area were evaluated. Comparing with field measured values, the prediction accuracy was up to80%. Crown density layers of3stages were generated then, for the analysis on the dynamic changes of forest canopy density. The results indicated that forest crown density in Aanshan showed a trend of "U-shaped" during1995to2006, with a best level in1995.
(3) Based on the data of aerial photography, with field inventory and visual plot materials, forest structural indexes were extracted by watershed algorithm. The extraction of tree number, canopy coverage, crown width, and location for single tree were analyzed and discussed then. Results showed the forest structural indexes can be extracted effectively, and it could be applied in forest survey for open forestland or urban area.
(4) Taking TM images at3stages in1995,2000and2006,as main data source, combined with the national forest inventory data, forest stock volume was estimated by stepwise regression method. The space-time dynamic progress of stock estimations for3stages were analysed respectively. Results showed a growth trend for forest stock volume from1995to2006, with estimations followed by688.11×105m3、755.86×105m3and1042.67×105m3. Comparing estimated volume in2006with sub-comparment statistic result of that in2007from forest resource inventory planning and design, we found no significant difference between them (13.79%), which suggested the model can ideally estimate volume, with high accuracy (above85%).
(5) Based on the national forest inventory data, growth models for height, DBH (diameter at breast height) and volume were finished by site classification method. All of that growth models were used for the updates of tree measured factors in forest resources survey. Meanwhile, prediction models for volume, growth and consumption at stand level were bulited with some related factors as age, height, DBH crown density, etc. Results showed that archival data of forest resources can be updated based on growth model, which would helped for the annual monitoring of forest resources.
(6). Our research improved the technology of forest resources monitoring based on Multi-data Source, and built a city level system for the annual monitoring of forest resources, which had been applicated successfully in Anshan. With the demonstration application, we learned the dynamic of forest resource from2007to2009, and also provided reliable basis and construction suggestions for forestry development and eco-environment planning.
(7) Our study promoted to comprehensive utilization of multi-source, multi-level remote sensing data. We studied quantitative extraction of forest area by multilevel remote sensing method based on PPS sampling, and combined with field survey data. Focused on dynamic monitoring of forest resource, we took remote sensing technology and spatial information technology as main methods for gathering data. And then, a system for the annual monitoring of forest resources in Anshan was built by series methods as remote sensing macroscopic monitoring&field investigation, quantitative analysis&qualitative analysis, model estimation&field measurements. This study improved the urgently need technology to monitor forest resource dynamic, and successfully applied to Anshan annual monitoring of forest resources dynamic. In conclusion, our research filled the blank of the municipal annual monitoring technology of forest resources. The techonology could greatly improve the ability of forest resources monitoring in Anshan, reduce field investigation cost, and met the need of municipal survey, supervision and enforcement of forest resources.
引文
[1]包盈智,袁凯先,赵宪文,等.森林资源二类调查中蓄积量遥感估测方法的应用实例[J].林业学研究,1996,9(3)：234-238.
[2]鲍云飞,曹春香,倪文俭,李增元,李小文.基于地基LiDAR的森林单木参数提取方法研究[J].第七届中国遥感大会摘要集.2010.
[3]曹春香,徐敏,何祺胜,张颢.多源遥感数据应用于森林健康研究的趋势[J],.遥感学报2009,13(s1)：401-407.
[4]曹春香,鲍云飞等.Retrieval of forest canopy attributes based on a geometric-optical model using airborne LiDAR and optical remote-sensing data. [J]. International Journal of Remote Sensing.2012,33(3):692-709.
[5]陈昌,黄纯忠.贵州省平坝县森林资源动态分析及评价[J].黑龙江生态工程职业学院学报,2008,21(5),25-26.
[6]陈楚,关泽群,张鹏林,等.利用RS和GIS的森林蓄积量偏最小二乘估测研究[J].湖北林业科技,2004(4)：25-28.
[7]陈瑞晶,明海军,王君女.内蒙古大兴安岭林区森林结构特征分析[J].内蒙古林业调查设计,2008,31(1),56-58.
[8]陈永富,杨彦臣,张怀清,杨秀森,李大江.海南岛热带天然山地雨林立地质量评价研究[J].林业科学研究,2000,13(2)：134-140.
[9]党永峰.遥感技术在森林资源连续清查中的应用——以利用遥感技术分析森林植被、地类的动态变化为例[J].林业资源管理.2004,(6)：63-65.
[10]范凤云.基于机载LIDAR和ALOSPALSAR全极化SAR数据的山区森林蓄积量估测方法研究[D].北京：中国林业科技研究院,2010
[11]冯仲科,杨伯钢,罗旭,等.应用LIDAR技术预测林分蓄积量[J].北京林业大学学报,2007(S2)：45-50.
[12]付晓,王雪军,韩爱惠,等.纸浆源地桉树资源遥感动态监测与分析[J].北京林业大学学报,2008,30(s3)：89-93.
[13]国红,陆元昌,洪玲霞等.基于RS与GIS的西峡县森林资源时空动态变化[J].南京林业大报.2005,29(4)：53-56.
[14]国家林业局(全国林业发展区划办公室).中国林业发展区划(综合篇)[M].北京：中国林业出版社,2011：4-36.
[15]国家林业局.国家森林资源连续清查主要技术规定[S].2004：23-26.
[16]黄建文.遥感及GIS技术在森林资源信息更新中的应用[J].林业资源管理.2002,(2)：59-62.
[17]黄平,杨燕琼,侯长谋.基于RS、GIS的杉木林分蓄积量判读模型研究[J].中南林业调查规划,2003,22(1)：25-27.
[18]侯长谋,杨燕琼,黄平,等.基于RS、GIS的马尾松林分蓄积量判读模型研究[J].林业资源管理[J],2002,(5)：55-57.
[19]侯长谋,张倩媚,彭少麟,等.3S技术在广东省森林资源动态监测中的应用[J].中南林业调查规划.2002,21(2)：44-45.
[20]贾永红,李德仁,孙家柄.多源遥感影像数据融合[J].遥感技术与应用,2000,15(1)：41-44.
[21]季碧勇,陶吉兴,王文武.基于连续清查固定样地生物量的立地质量评价[J].西南林业大学学报,2012,32(4)：45-50.
[22]金大刚,李明..1997-2005年广西森林资源变化动态评价[J].广西林业科学,2007,36(4),181-186.
[23]金久宏.萧山区森林结构优化分析[J].林业科技情报,2009,41(1),3-5.
[24]李锦烨.森林资源结构动态分析及其调整的研究.学位论文.2010
[25]李崇贵,赵宪文,李春干.森林蓄积量遥感估测理论与实现[M].北京：科学出版社,2006
[26]李崇贵.用非线性理论研究以“3S”为基础的森林蓄积量估测[D].北京：中国林业科学研究院,2001.
[27]李春于.基于卫星遥感数据空中抽样的大尺度森林资源动态监测[J]..林业资源管理.2009,(2)：106-110.
[28]李海奎,雷渊才.中国森林植被生物量和碳储量评估[M].北京：中国林业出版社.2010：5-25.
[29]李贞猷.福建省森林资源开发利用动态分析及对策措施[J]..林业勘察设计,2001,(2),16-19.
[30]李小文,王锦地.植被光学遥感模型和植被结构参数化[M].北京：科学出版社,1995.
[31]吕勇,朱光玉,罗立平,周根苗,邓成.基于不同立地因子的杉木与马尾松地位指数相关模型[J].中南林业科技大学学报,2007,27：81-84.
[32]马秀芳,林媚珍.广东省森林资源动态分析及发展方向[J]..自然地理学与生态建设,2006,7,353-358.
[33]孟宪宇.测树学(第3版)[M].北京：中国林业出版社,2006：98-112.
[34]舒清态,唐守正.国际森林资源监测的现状与发展趋势[J].世界林业研究,2006,18(3)：33-37.
[35]谭炳香.高光谱遥感森林类型识别及其郁闭度定量估测研究[D].北京：中国林业科学研究院,2006.
[36]唐守正,郎奎建,李海奎.统计和生物数学模型计算(ForStat教程)[M].北京：科学出版社,2009：115-121.
[37]滕维超,万文生,王凌晖.森林立地分类与质量评价研究进展[J].广西农业科学,2009,40(8)：1110-1114.
[38]王臣立,牛铮,郭治兴,等RadarsatSAR的森林生物物理参数信号响应及其蓄积量估测[J].国土资源遥感,2005,(2)：24-28.
[39]王维芳,郎奎建.方山林场森林资源结构动态分析与经营对策[J].东北林业大学学报[J],2004,32(3)：4-7.
[40]王雪军,韩爱惠,黄国胜,张煜星,丁晓强.国家级森林资源遥感监测业务运行系统的设计[J].林业资源管理,2005,(2)：70-74.
[41]肖化顺.森林资源监测中林业3S技术的应用现状与展望[J].林业资源管理.2004,(2)：53-58.
[42]许东,代力民,邵国凡,等.基于RS、GIS及k-近邻法的森林蓄积量估测[J].辽宁工程技术大学学报(自然科学版),2008,2(27)：195-197.
[43]杨永恬,李增元,陈尔学,等.基于ALOSPALSAR数据的森林蓄积量估测技术研究[J].林业资源管理,2010(2)：13-117.
[44]杨永恬.基于多源遥感数据的森林蓄积量估测方法研究[D].北京：中国林业科学研究院,2010.
[45]姚静,武文波,康停军.多源遥感影像数据融合技术探讨[J].测绘与空间地理信息,2008,31(1)：29-31.
[46]袁凯先,陈玉山,包盈智,等.森林蓄积量的遥感估测[J].林业资源管理,1996(3)：13-17.
[47]余坤勇,林芳,刘健,等.基于RS的闽江流域马尾松林分蓄积量估测模型研究[J].2006,33(1)：16-23.
[48]游先祥.遥感原理及在资源环境中的应用[M].北京：中国林业出版社,2003.
[49]张远,杨彬,殷鸣放等.山区森林资源遥感动态监测的研宄[J].福建林业科技.2005,32(1)：30-33.
[50]张煜星,王祝雄,等编著.遥感技术在森林资源清查中的应用研究[M].北京：中国林业出版社,2007.
[51]张彦林.基于3S技术的山东省森林蓄积量定量估测研究[D].北京：北京林业大学,2008.
[52]张彦林BOX-COX变换在遥感数据建模中的应用[J].东北林业大学学报,2010,8(38)：120-122.
[53]张友静,方有清.K-T变换的林学意义及其在森林蓄积量估算中的应用[J].环境遥感,1992,7(3)：163-171.
[54]赵宪文.林业遥感定量估测[M].北京：中国林业出版社,1996.
[55]赵宪文,李崇贵,斯林,等.森林资源遥感估测的重要进展[J].中国工程科学,2001,3(8)：15-24.
[56]郑小贤.瑞典、瑞士和芬兰多目的森林环境监测[J].世界林业研究,1997,(2)：58-64.
[57]周成贵.云南永平县森林资源动态分析及发展建议[J].林业科技情报,2008,40(3),5-7.
[58]朱海珍,庞勇,杨飞,等.基于ENVISATASTAR数据的森林蓄积量估测研究[J].地理与地理信息科学,2007,23(2)：51-55.
[59]Arun D Kulkarni. Neural-Fuzzy models for multispectral image analysis[J]. Applied Intelligence,1998,8 (2):173-187.
[60]Binaghi E,Gallo I and Pepe M. A neural adaptive model for feature extraction and recognition in high resolution remote sensing imagery[J]. International Journal of Remote Sensing,2003,24 (20): 3947-3959.
[61]Cole D W, Van Miegroet H. Chronosequences:a technique to assess ecosystem dynamics. In:Research strategies for long-term site productivity. Proceedings, IEA/BE A3 Workshop,Seattle,WA, August 1988.(Eds.)
[62]Daya Nand Pant Susane M.E, GrotenPartha sarthi Roy,2000.FOREST VEGETATION/ LANDUSE CHANGE DETECTION AND IMPACTASSESSMENT IN PART OF WESTERN HIMALAYA, International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam. p1111-p1118.
[63]Ehmana J L et al.2002, An integrated GIS andmodeling approach for assessing the transient response of forests of the southern Great Lakes region to a doubled COZ climate. Forest Ecology and Management,155:237-255.
[64]Friedl M A,Brodley C E and Strahler A H. Maximizing land cover classification accuracies produced by decision trees at continental to global scales[J]. IEEE Transactionson Geosciencesand Remote Sensing,1999,37 (2):969-979.
[65]Fugui Wang-Y. Jun Xu,2010.Comparison of remote sensing change detectiontechniques for assessing hurricane damage to forests,Environ Monit Assess.162:311-326,2010.
[66]Hansen M C,Defries R S,Townshend J R G,et al. Global land cover classification at 1 km spatial resolution using a classification tree approach[J]. International Journal of Remote Sensing,2000,21 (6):1331-1364.
[67]J. Bohlin, J. Wallerman, H.Olsson, J. E. S. Fransson,2012. species-specific forest variable estinmation using non-parametric modeling of multi-spectral photogrammetric point cloud data.International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B8,2012,XXII ISPRS Congress,25.Melbourne, Australia.
[68]Jordan C F. Derivation of leaf area index from quality of light on the forest floor[J]. Ecology,1969,50:663-666.
[69]Knight C L, et al..1994. Expansion of gallery forest on Konza Prairie Research Natural Area Kansas..Landscape Ecology,9(2):117-125.
[70]LU D, MAUSEL P,BRONDIZIO E,et al.2004. Change detection techniques.International Journal of Remote Sensing.25(12):2365-2407.
[71]Luis Marcelo Tavares de Carvalho,Jan G. P. W. Clevers,Steven M. de Jong.2005.Andrew K. Skidmore Forestry database updating based on remote sensing change detection,Anais XII Simposio Brasileiro de Sensoriamento Remoto, Goiania, Brasil,16-21, INPE, p.465-472.
[72]Luis W. Rivera T. Mitchell Aide.1998.Forest recovery in the karsts region of Puerto Rico. Forest Ecology and Management,(108):63-75.
[73]Michael R.Saunders.2006,Dynamics of forest structure under different silvicultural regimes in the Acadian forest. Submitted in Partial Fulfillment of the Requirements of the Degree, of Doctor of Philosophy.
[74]MUCHONEY D M HAACK B N.1994.Change detection for monitoring forest defoliatio. Photogrammetric Engineering and Remote Sensing.60:1243-1251.
[75]Nicholas N S., Zedaker S. M. Espected stand behavior:Site quality estimation for southern appalachian red spruce. Forest. Ecology. Management.,1992,47:39-50.
[76]Okin G S and Gillete D A. Distribution of vegetation in wind-dominated landscapes:Implications for wind erosion modeling and landscape processes [J]. Journal of Geophysical Research,2001,106 (D9):9673-9683
[77]Paola J D and Schowengerdt R A. A detailed comparison of back-propagation neural network and maximum-likelihood classifiers for urban land use classification[J]. IEEE Transactions Geosciences Remote Sensing,1995,33 (4):981-996
[78]Pu Ruiliang,Peng Gong. Wavelet Transform Applied to EO-1 Hyperspectral Data for Forest LAI and Crown Closure Mapping [J]. Remote Sensing of Environment,2004,91:212-224
[79]ROGAN J,FRANKLIN J.ROBERTS D A.2002. A comparison of methods for monitoringmultitemporal vegetation change usingThematic Mapper imagery.Remote Sensing of Enviroment.80(1):143-156.
[80]Simard M,Saatchi S S.Grandi G D,et al. The use of decision tree and multiscale texture for classification of JERS-1SAR data over tropical forest[J]. IEEE Transactions on Geosciences and Remote Sensing,2000,38 (5):2310-2321
[81]Siddiqui M N et al.2004,Monitoring changes in riverine forests of Sindh-Pakistan using remote sensing and GIS techniques. Advances in Space Research,33:333-337.
[82]Singh A.1986. Change detection in the tropical forest environment of northeastern India using Landsat. Remote Sensing and Tropical Land Management. pp.237-254.
[83]Stout B B, Shumway D L. Site quality estimation using height and diameter. Forest Science, 1982,25:639-645.
[84]Vito Alberga.2009. Similarity Measures of Remotely Sensed Multi-Sensor Imagesfor Change Detection Applications.Remote Sens.1,122-143.
[85]Weiskittel, A. R., Hann, D. W., Kershaw, J. A. and Vanclay, J. K. (2011) Forest Site Evaluation, in Forest Growth and Yield Modeling, John Wiley & Sons, Ltd, Chichester, UK. doi: 10.1002/9781119998518.ch3
[2]鲍云飞,曹春香,倪文俭,李增元,李小文.基于地基LiDAR的森林单木参数提取方法研究[J].第七届中国遥感大会摘要集.2010.
[3]曹春香,徐敏,何祺胜,张颢.多源遥感数据应用于森林健康研究的趋势[J],.遥感学报2009,13(s1)：401-407.
[4]曹春香,鲍云飞等.Retrieval of forest canopy attributes based on a geometric-optical model using airborne LiDAR and optical remote-sensing data. [J]. International Journal of Remote Sensing.2012,33(3):692-709.
[5]陈昌,黄纯忠.贵州省平坝县森林资源动态分析及评价[J].黑龙江生态工程职业学院学报,2008,21(5),25-26.
[6]陈楚,关泽群,张鹏林,等.利用RS和GIS的森林蓄积量偏最小二乘估测研究[J].湖北林业科技,2004(4)：25-28.
[7]陈瑞晶,明海军,王君女.内蒙古大兴安岭林区森林结构特征分析[J].内蒙古林业调查设计,2008,31(1),56-58.
[8]陈永富,杨彦臣,张怀清,杨秀森,李大江.海南岛热带天然山地雨林立地质量评价研究[J].林业科学研究,2000,13(2)：134-140.
[9]党永峰.遥感技术在森林资源连续清查中的应用——以利用遥感技术分析森林植被、地类的动态变化为例[J].林业资源管理.2004,(6)：63-65.
[10]范凤云.基于机载LIDAR和ALOSPALSAR全极化SAR数据的山区森林蓄积量估测方法研究[D].北京：中国林业科技研究院,2010
[11]冯仲科,杨伯钢,罗旭,等.应用LIDAR技术预测林分蓄积量[J].北京林业大学学报,2007(S2)：45-50.
[12]付晓,王雪军,韩爱惠,等.纸浆源地桉树资源遥感动态监测与分析[J].北京林业大学学报,2008,30(s3)：89-93.
[13]国红,陆元昌,洪玲霞等.基于RS与GIS的西峡县森林资源时空动态变化[J].南京林业大报.2005,29(4)：53-56.
[14]国家林业局(全国林业发展区划办公室).中国林业发展区划(综合篇)[M].北京：中国林业出版社,2011：4-36.
[15]国家林业局.国家森林资源连续清查主要技术规定[S].2004：23-26.
[16]黄建文.遥感及GIS技术在森林资源信息更新中的应用[J].林业资源管理.2002,(2)：59-62.
[17]黄平,杨燕琼,侯长谋.基于RS、GIS的杉木林分蓄积量判读模型研究[J].中南林业调查规划,2003,22(1)：25-27.
[18]侯长谋,杨燕琼,黄平,等.基于RS、GIS的马尾松林分蓄积量判读模型研究[J].林业资源管理[J],2002,(5)：55-57.
[19]侯长谋,张倩媚,彭少麟,等.3S技术在广东省森林资源动态监测中的应用[J].中南林业调查规划.2002,21(2)：44-45.
[20]贾永红,李德仁,孙家柄.多源遥感影像数据融合[J].遥感技术与应用,2000,15(1)：41-44.
[21]季碧勇,陶吉兴,王文武.基于连续清查固定样地生物量的立地质量评价[J].西南林业大学学报,2012,32(4)：45-50.
[22]金大刚,李明..1997-2005年广西森林资源变化动态评价[J].广西林业科学,2007,36(4),181-186.
[23]金久宏.萧山区森林结构优化分析[J].林业科技情报,2009,41(1),3-5.
[24]李锦烨.森林资源结构动态分析及其调整的研究.学位论文.2010
[25]李崇贵,赵宪文,李春干.森林蓄积量遥感估测理论与实现[M].北京：科学出版社,2006
[26]李崇贵.用非线性理论研究以“3S”为基础的森林蓄积量估测[D].北京：中国林业科学研究院,2001.
[27]李春于.基于卫星遥感数据空中抽样的大尺度森林资源动态监测[J]..林业资源管理.2009,(2)：106-110.
[28]李海奎,雷渊才.中国森林植被生物量和碳储量评估[M].北京：中国林业出版社.2010：5-25.
[29]李贞猷.福建省森林资源开发利用动态分析及对策措施[J]..林业勘察设计,2001,(2),16-19.
[30]李小文,王锦地.植被光学遥感模型和植被结构参数化[M].北京：科学出版社,1995.
[31]吕勇,朱光玉,罗立平,周根苗,邓成.基于不同立地因子的杉木与马尾松地位指数相关模型[J].中南林业科技大学学报,2007,27：81-84.
[32]马秀芳,林媚珍.广东省森林资源动态分析及发展方向[J]..自然地理学与生态建设,2006,7,353-358.
[33]孟宪宇.测树学(第3版)[M].北京：中国林业出版社,2006：98-112.
[34]舒清态,唐守正.国际森林资源监测的现状与发展趋势[J].世界林业研究,2006,18(3)：33-37.
[35]谭炳香.高光谱遥感森林类型识别及其郁闭度定量估测研究[D].北京：中国林业科学研究院,2006.
[36]唐守正,郎奎建,李海奎.统计和生物数学模型计算(ForStat教程)[M].北京：科学出版社,2009：115-121.
[37]滕维超,万文生,王凌晖.森林立地分类与质量评价研究进展[J].广西农业科学,2009,40(8)：1110-1114.
[38]王臣立,牛铮,郭治兴,等RadarsatSAR的森林生物物理参数信号响应及其蓄积量估测[J].国土资源遥感,2005,(2)：24-28.
[39]王维芳,郎奎建.方山林场森林资源结构动态分析与经营对策[J].东北林业大学学报[J],2004,32(3)：4-7.
[40]王雪军,韩爱惠,黄国胜,张煜星,丁晓强.国家级森林资源遥感监测业务运行系统的设计[J].林业资源管理,2005,(2)：70-74.
[41]肖化顺.森林资源监测中林业3S技术的应用现状与展望[J].林业资源管理.2004,(2)：53-58.
[42]许东,代力民,邵国凡,等.基于RS、GIS及k-近邻法的森林蓄积量估测[J].辽宁工程技术大学学报(自然科学版),2008,2(27)：195-197.
[43]杨永恬,李增元,陈尔学,等.基于ALOSPALSAR数据的森林蓄积量估测技术研究[J].林业资源管理,2010(2)：13-117.
[44]杨永恬.基于多源遥感数据的森林蓄积量估测方法研究[D].北京：中国林业科学研究院,2010.
[45]姚静,武文波,康停军.多源遥感影像数据融合技术探讨[J].测绘与空间地理信息,2008,31(1)：29-31.
[46]袁凯先,陈玉山,包盈智,等.森林蓄积量的遥感估测[J].林业资源管理,1996(3)：13-17.
[47]余坤勇,林芳,刘健,等.基于RS的闽江流域马尾松林分蓄积量估测模型研究[J].2006,33(1)：16-23.
[48]游先祥.遥感原理及在资源环境中的应用[M].北京：中国林业出版社,2003.
[49]张远,杨彬,殷鸣放等.山区森林资源遥感动态监测的研宄[J].福建林业科技.2005,32(1)：30-33.
[50]张煜星,王祝雄,等编著.遥感技术在森林资源清查中的应用研究[M].北京：中国林业出版社,2007.
[51]张彦林.基于3S技术的山东省森林蓄积量定量估测研究[D].北京：北京林业大学,2008.
[52]张彦林BOX-COX变换在遥感数据建模中的应用[J].东北林业大学学报,2010,8(38)：120-122.
[53]张友静,方有清.K-T变换的林学意义及其在森林蓄积量估算中的应用[J].环境遥感,1992,7(3)：163-171.
[54]赵宪文.林业遥感定量估测[M].北京：中国林业出版社,1996.
[55]赵宪文,李崇贵,斯林,等.森林资源遥感估测的重要进展[J].中国工程科学,2001,3(8)：15-24.
[56]郑小贤.瑞典、瑞士和芬兰多目的森林环境监测[J].世界林业研究,1997,(2)：58-64.
[57]周成贵.云南永平县森林资源动态分析及发展建议[J].林业科技情报,2008,40(3),5-7.
[58]朱海珍,庞勇,杨飞,等.基于ENVISATASTAR数据的森林蓄积量估测研究[J].地理与地理信息科学,2007,23(2)：51-55.
[59]Arun D Kulkarni. Neural-Fuzzy models for multispectral image analysis[J]. Applied Intelligence,1998,8 (2):173-187.
[60]Binaghi E,Gallo I and Pepe M. A neural adaptive model for feature extraction and recognition in high resolution remote sensing imagery[J]. International Journal of Remote Sensing,2003,24 (20): 3947-3959.
[61]Cole D W, Van Miegroet H. Chronosequences:a technique to assess ecosystem dynamics. In:Research strategies for long-term site productivity. Proceedings, IEA/BE A3 Workshop,Seattle,WA, August 1988.(Eds.)
[62]Daya Nand Pant Susane M.E, GrotenPartha sarthi Roy,2000.FOREST VEGETATION/ LANDUSE CHANGE DETECTION AND IMPACTASSESSMENT IN PART OF WESTERN HIMALAYA, International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam. p1111-p1118.
[63]Ehmana J L et al.2002, An integrated GIS andmodeling approach for assessing the transient response of forests of the southern Great Lakes region to a doubled COZ climate. Forest Ecology and Management,155:237-255.
[64]Friedl M A,Brodley C E and Strahler A H. Maximizing land cover classification accuracies produced by decision trees at continental to global scales[J]. IEEE Transactionson Geosciencesand Remote Sensing,1999,37 (2):969-979.
[65]Fugui Wang-Y. Jun Xu,2010.Comparison of remote sensing change detectiontechniques for assessing hurricane damage to forests,Environ Monit Assess.162:311-326,2010.
[66]Hansen M C,Defries R S,Townshend J R G,et al. Global land cover classification at 1 km spatial resolution using a classification tree approach[J]. International Journal of Remote Sensing,2000,21 (6):1331-1364.
[67]J. Bohlin, J. Wallerman, H.Olsson, J. E. S. Fransson,2012. species-specific forest variable estinmation using non-parametric modeling of multi-spectral photogrammetric point cloud data.International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B8,2012,XXII ISPRS Congress,25.Melbourne, Australia.
[68]Jordan C F. Derivation of leaf area index from quality of light on the forest floor[J]. Ecology,1969,50:663-666.
[69]Knight C L, et al..1994. Expansion of gallery forest on Konza Prairie Research Natural Area Kansas..Landscape Ecology,9(2):117-125.
[70]LU D, MAUSEL P,BRONDIZIO E,et al.2004. Change detection techniques.International Journal of Remote Sensing.25(12):2365-2407.
[71]Luis Marcelo Tavares de Carvalho,Jan G. P. W. Clevers,Steven M. de Jong.2005.Andrew K. Skidmore Forestry database updating based on remote sensing change detection,Anais XII Simposio Brasileiro de Sensoriamento Remoto, Goiania, Brasil,16-21, INPE, p.465-472.
[72]Luis W. Rivera T. Mitchell Aide.1998.Forest recovery in the karsts region of Puerto Rico. Forest Ecology and Management,(108):63-75.
[73]Michael R.Saunders.2006,Dynamics of forest structure under different silvicultural regimes in the Acadian forest. Submitted in Partial Fulfillment of the Requirements of the Degree, of Doctor of Philosophy.
[74]MUCHONEY D M HAACK B N.1994.Change detection for monitoring forest defoliatio. Photogrammetric Engineering and Remote Sensing.60:1243-1251.
[75]Nicholas N S., Zedaker S. M. Espected stand behavior:Site quality estimation for southern appalachian red spruce. Forest. Ecology. Management.,1992,47:39-50.
[76]Okin G S and Gillete D A. Distribution of vegetation in wind-dominated landscapes:Implications for wind erosion modeling and landscape processes [J]. Journal of Geophysical Research,2001,106 (D9):9673-9683
[77]Paola J D and Schowengerdt R A. A detailed comparison of back-propagation neural network and maximum-likelihood classifiers for urban land use classification[J]. IEEE Transactions Geosciences Remote Sensing,1995,33 (4):981-996
[78]Pu Ruiliang,Peng Gong. Wavelet Transform Applied to EO-1 Hyperspectral Data for Forest LAI and Crown Closure Mapping [J]. Remote Sensing of Environment,2004,91:212-224
[79]ROGAN J,FRANKLIN J.ROBERTS D A.2002. A comparison of methods for monitoringmultitemporal vegetation change usingThematic Mapper imagery.Remote Sensing of Enviroment.80(1):143-156.
[80]Simard M,Saatchi S S.Grandi G D,et al. The use of decision tree and multiscale texture for classification of JERS-1SAR data over tropical forest[J]. IEEE Transactions on Geosciences and Remote Sensing,2000,38 (5):2310-2321
[81]Siddiqui M N et al.2004,Monitoring changes in riverine forests of Sindh-Pakistan using remote sensing and GIS techniques. Advances in Space Research,33:333-337.
[82]Singh A.1986. Change detection in the tropical forest environment of northeastern India using Landsat. Remote Sensing and Tropical Land Management. pp.237-254.
[83]Stout B B, Shumway D L. Site quality estimation using height and diameter. Forest Science, 1982,25:639-645.
[84]Vito Alberga.2009. Similarity Measures of Remotely Sensed Multi-Sensor Imagesfor Change Detection Applications.Remote Sens.1,122-143.
[85]Weiskittel, A. R., Hann, D. W., Kershaw, J. A. and Vanclay, J. K. (2011) Forest Site Evaluation, in Forest Growth and Yield Modeling, John Wiley & Sons, Ltd, Chichester, UK. doi: 10.1002/9781119998518.ch3