The Terrestrial Laser Scanning (TLS) has the characteristics of non-destructive measurement, and its point cloud information acquired from TLS allows obtaining accurate forest attributes through visualization and automated extraction. The purposes of this study include using TLS to obtain point cloud information and discuss different methods to extract Diameter at breast height (DBH) and height (H) within small plot, and we also discuss the impact of point cloud generation in different resolution and station number settings. The results showed that the DBH and H measured by TLS were not significantly different from the measured values in the manual extraction and semi-automatic extraction detection methods. In addition, when the scanning resolution is 1/4(43.70 MPts) and 1/2(174.80 MPts), there is no significant difference between the measured DBH, H and the measured value, and the accuracy does not increase when the resolution is 1/2(174.80 MPts). Therefore, we can obtain the accurate DBH at the scanning resolution of 1/4(43.70 MPts); and the number of scanning stations is positively correlated with the DBH and H extraction accuracy. Therefore, the number of stations should be more than four stations.

The FOMROSAT-8 (FS-8) program integrates two vegetation red edge (VRE) bands into its two-meter pushbroom multi-spectral sensors. Time-Delay-Integration (TDI) technique is utilized to improve Signal-to-Noise Ratio (SNR). Along with other applications, these two VRE bands are fruitful for the retrieval of chlorophyll contents of vegetation canopies. Without TDI, high spatial resolution and narrow bandwidth result in poor SNR which weakens the applicability of VRE bands. In order to obtain reasonable number of TDI stages setting for VRE bands, firstly the spectral responses of chlorophyll contents are simulated to obtain the noise affected reflectances observed by different number of TDI stages setting. Then, acceptable accuracy of retrieved chlorophyll contents and the corresponding number of TDI stages setting are suggested as a reference in the design of FS-8.

Green Water Footprint (GWF) is referred to total rainwater evapotranspiration (ET) plus the amount of water incorporating in a product. As the water incorporated into the crop is about 0.1~1% of the evapotranspiration volume, GWF is normally referred to as ET volume. Together with irrigation water withdrawn from ground or surface water, they are the main indicators of contribution of water usage introduced in agricultural products. Therefore in order to comprehensively understand the total amount of water used during crop growing stage, it is essential to identify the amount of and Green water. Based on the high accuracy, wide coverage and long-term extraction of rainwater evapotranspiration, a MODIS Global Terrestrial Evapotranspiration Data Set (MOD16) is applied to estimate GWF. Although the MOD16 product offers aforementioned advantages, we still encountered spatial and temporal inaccessibility. Therefore, this research aimed to overcome the drawbacks and develop a regression method considering multiple variables to improve the performance of GWF estimation. Through the experiments implemented in Taichung and Taitung, it was successfully demonstrated that the spatial and temporal accessibility of GWF of rice paddies calculated based on MOD16 data was improved significantly.

In the past, it draws the great attraction of using Data Mining approaches and geostatistics analysis through remote sensing data which are well-accepted. The goal of this study is decided to extract the core spectral information through hyperspectral vs. multispectral imaging. More specifically, the paddy-field remote sensing image is applied with a supervised learning linear discriminant analysis and unsupervised learning density-based clustering algorithm in this study. The pre-processing is used the Principal Component Analysis (PCA) to design parallel study for four case studies: (1) hyper-spectrum versus multi-spectrum with linear discriminant analysis (2) hyper-spectrum versus multi-spectrum density-based clustering algorithm (3) hyper-spectrum versus multispectrum principal component analysis + linear discriminant analysis (4) hyper-spectrum versus multi-spectrum by principal component + density-based clustering algorithm. The DBSCAN with hyper-spectrum image data has an overall accuracy rate of 86.85% which is higher than those of DBSCAN with multi-spectrum (79.45%). The results are presented by error matrix (accuracy rate) and the thematic maps are drawn.