Leaf area index (LAI) is a key variable in many land surface models that involve energy and mass exchange between vegetation and the environment. In recent years, extracting vegetation structure parameters from digital photography becomes a widely used indirect method to estimate LAI for its simplicity and ease of use. A Leaf Area Index Sensor (LAIS) system was developed to continuously monitor the growth of crops in several sampling points in Huailai, China. The system applies 3G/WIFI communication technology to remotely collect crop photos in real-time. Then the crop photos are automatically processed and LAI is estimated based on the improved leaf area index of Lang and Xiang (LAILX) algorithm in LAIS. The objective of this study is to primarily verify the LAI estimated from LAIS (Lphoto) through comparing them with the destructive green LAI (Ldest). Ldest was measured across the growing season ntil maximum canopy development while plants are still green. The preliminary verification shows that Lphoto corresponds well with the Ldest (R2=0.975). In general, LAI could be accurately estimated with LAIS and its LAI shows high consistency compared with the destructive green LAI. The continuous LAI measurement obtained from LAIS could be used for the validation of remote sensing LAI products.
The Wireless Sensor Networks of Coarse-resolution Pixel Parameters (CPP-WSN) was established to monitor the heterogeneity of coarse spatial resolution pixel, with consideration of different categories of land surface parameters in Huailai, Hebei province, China (40.349°N, 115.785°E). The observation network of radiation parameters (RadNet) in CPP-WSN was developed for multi-band radiation measurement and consisted of 6 nodes covering 2km*2km area to capture its heterogeneity. Each node employed four sensors to observe the five radiation parameters. The number and location of nodes in RadNet were determined through the representativeness-based sampling method. Thus, the RadNet is a distributed observation system with nodes work synchronously and measurements used together. <p> </p>The intercomparison experiment for RadNet is necessary and was conducted in Huailai Remote Sensing Station from 5th Aug to 10th Aug in 2012. Time series observations from various sensors were collected and analyzed. The maximum relative differences among sensors of UVR, SWR, LWR, PAR, and LST are 4.83%, 5.3%, 3.71%, 11%, and 0.54%, respectively. Sensor/parameter differences indeed exist and are considerable large for PAR, SWR, UVR, and LWR, which cannot be ignored. The linear normalization and quadratic polynomial normalization perform similar for CUV5/UVR, PQS1/PAR, CNR4/SWR, and SI-111/LST. As for CNR4/LWR, quadratic polynomial normalization show higher accuracy than linear normalization, especially in node2, node4, and node5. Thus, the LWR measured by CNR4 is proved to be nonlinear, and should be normalized with quadratic polynomial coefficients for higher precision.
The long term record of remote sensing product shows the land surface parameters with spatial and temporal change to
support regional and global scientific research widely. Remote sensing product with different sensors and different
algorithms is necessary to be validated to ensure the high quality remote sensing product. Investigation about the remote
sensing product validation shows that it is a complex processing both the quality of in-situ data requirement and method
of precision assessment. A comprehensive validation should be needed with long time series and multiple land surface
types. So a system named as land surface remote sensing product is designed in this paper to assess the uncertainty
information of the remote sensing products based on a amount of in situ data and the validation techniques.
The designed validation system platform consists of three parts: Validation databases Precision analysis subsystem,
Inter-external interface of system. These three parts are built by some essential service modules, such as Data-Read
service modules, Data-Insert service modules, Data-Associated service modules, Precision-Analysis service modules,
Scale-Change service modules and so on. To run the validation system platform, users could order these service modules
and choreograph them by the user interactive and then compete the validation tasks of remote sensing products (such as
LAI ,ALBEDO ,VI etc.) .
Taking SOA-based architecture as the framework of this system. The benefit of this architecture is the good service
modules which could be independent of any development environment by standards such as the Web-Service
Description Language(WSDL). The standard language: C++ and java will used as the primary programming language to
create service modules.
One of the key land surface parameter, albedo, is selected as an example of the system application. It is illustrated that
the LAPVAS has a good performance to implement the land surface remote sensing product validation.