A Study on Forest Classification and Dynamic Monitoring by Using Remote Sensing Information[J]. Journal of Remote Sensing, 1995, (2). DOI: 10.11834/jrs.1995014.
In the investigation of forest resources and dvnamic monitoring
it isgenerally required to distinguish tree species.If the species are complicated
it iSalso required to distinguish the species groups with similar biological characteristics.In drawing the maps it is required to have great geometric accuracy.The mediumscale image with distinct type border(for example
the image of 1/25000)can becomposed and drawn directly.To meet the needs of the investigation of forest re-sources and dynamic monitoring
we conducted the study on information compound. It includ 5 the study on the practical techniques of remote sensing information com-pound
remote sensing and non-remote-sensing information compounds
applicationof auxiliary information
compound method
etc.In experiments
we used the studiesof the satellite images of MSS and TM
the images of different temporal images ofTM
the compound of TM and SPOT
the quantitative analysis of the characteristicchanges of information before and after compound
etc.Thus provide basis for op-timum wave-band combination and the method of extracting characteristic informa-tion
and also provide the best satellite images of 1/25000 on more than 500 thou-sand hectares of and for the investigation of forest resources in the region of BaiLong Jiang River to conduct the second type of the investigation of forest resour-ces.In the analysis of dynamic monitoring
the following land type were mainlydetermined:cut-over land
forest wall and regeneration of cut-over land
etc.;theassessment Of accuracy of determination was also conducted.