The protection forest is distributed scattering among the island of Taiwan, its function in safeguarding the territory is very important. How to establish an efficient protection forest management policy is a main task for the manager. This study focused on how to use a new technique of remote sensing to research landscape structure of the protection forest within Kenting National Park. Multivariate analysis was used to group the 20 protection forest into 3 groups, and according to the landscape ecology theory to name them as diversity landscape structure, simple landscape structure and fragmentation landscape structure. Moreover we select NP (number of patches), ED (edge density) and SHDI (Shannon’s diversity indices) index to assess the disturbance of protection forest. The result showed that the region which protection forest impact seriously was that the human activity frequently or northeast monsoon attacked area.

The objective of this study was to estimate tree heights from point clouds scanned by a terrestrial laser scanner over a dense forest where the airborne lidar can not obtain enough ground return points. A 50m×50m sample plot was chosen in Yangminshan National Park where is located in northern Taiwan, and point clouds were then collected by utilizing Leica HDS3000 terrestrial laser scanner. Coordinates of each tree were surveyed by a total station and tree heights were estimated by subtract local minimum elevation from local maximum elevation by searching the point clouds within a specified radius at each surveyed tree location. True heights of trees in the sample plot were manually measured and were compared to the estimated tree heights. The result shows that the heights of dominate trees growing erectly can be estimated by simply subtracting the minimum elevation from maximum elevation, and the correctness of estimation are within 5% of the surveyed height, but results in other situations were not perform well. Other algorithm for point searching and tree height estimation must be then developed.

This investigation based on the characteristics of “Split-Merge-Shape” algorithm, utilizes a semi-automated strategy for building reconstruction. The original data used are the multi-overlapping aerial images. Since the SMS algorithm adopts incomplete roof-edges for building reconstruction, we proposed an interactive scheme that utilizes the image feature extraction and image matching techniques for non-hidden roof-edges extraction. First, an operator may manually locate the rough position of building corners in an image to obtain its approximate coordinates. Subsequently, the linear feature extraction and matching techniques were performed to extract precise location of roof-edges. Then, the 3D roof-edges can be calculated by means of photogrammetric space resection. Finally, we can adopt the SMS algorithm for building reconstruction. Experiment results show that the proposed scheme is effective provided that the building roof-edges are linear and high contrast with continuous height variation.

This study focuses on using techniques such as digital terrain model (DTM), geographic information system (GIS), and decision support system (DSS) for forestland ecosystem classification and suitability analysis of the Liukuei Experimental Forest of Taiwan Forestry Research Institute. The content includes the delineation of ecosystem units using DTM, the development of a hierarchical ecosystem classification system using GIS and multivariate statistical analysis, and the establishment of a forestland classification DSS with an application on site selection of a Taiwanese native species – Taiwania (Taiwania cryptomerioides). The results indicate that DTM is a fast, easy, feasible, and automatic approach for delineating ecosystem units of different spatial scales. The developed hierarchical ecosystem classification is a satisfactory scheme for Liukuei’s forestland classification because the developed scheme coincides with the terrain characteristics along a continuum. The established DSS can effectively and feasibly analyze forestland classification under different spatial scales. Meanwhile, the system can easily perform site selection for Taiwania. From the results, it is concluded that techniques such as DTM, GIS, and DSS are useful for forest managers in the reasonable planning of forestland classification and management practice. In addition, the ecosystem approaches obtained from the Experimental Forest can be extended to island-wide forestland classification in Taiwan.

This investigation is to integrate the Lidar data and high resolution satellite image for building detection. The proposed scheme includes preprocessing, segmentation, and classification. The preprocessing includes point cloud classification, rasterization, and spatial registration. We employ ground control points and rational polynomial coefficients in spatial registration. Thus, the two data sets are unified in the same object coordinate system. In the segmentation, we combine the raster form Lidar data and high resolution satellite orthoimages. The data with similar heights and spectral attributes are merged into a region. In the classification, we use the region-based decision classification to separate the building and non-building regions. The attributes considered in the classification include: elevation, spectral, texture and shape of regions. LIDAR data acquired by Leica ALS 40 and QuickBird satellite images were used in the validation. A 1/1000 scale topographic map was used as ground truth. The experiment results indicate that the building detection may exceed 89% of accuracy.

The purpose of monitoring forest landscape change is not only to understand the change of past and present landscape structure, but also to predict the future trend for providing the reference of forest management. This study applied Shannon diversity index, which was related to human disturbance, to qualify the landscape change of Lien-Hwa-Chih Experimental Forest between 1971 and 1998. The t-test based on Shannon diversity index was also used to examine the disturbance differences caused by unlawful cultivated forestland. In addition, the mathematical models such as Markov chain and binomial logit regression analysis were applied to predict the future overall landscape structure and to simulate the spatial distribution of unlawful cultivated forestland. Again, the t-test was applied to monitor the future effect of disturbance caused by unlawful cultivated forestland. The result indicated that, using Shannon t-test, the landscape change between 1971 and 1998 had no differences at the 5% significance level. That is, there was no significant differences about the disturbance caused by unlawful cultivated forestland. The prediction of landscape change in 2052, 2106, and 2160 using the Markov chain model indicated that the area of unlawful cultivated forestland was increased from original 0.90% in 1971 and 3.39% in 1998 to 7.85% in 2052, 12.08% in 2106, and 6.09% in 2160 ,respectively. As for the spatial distribution using binomial logit regression analysis, the result showed that the occurrence of unlawful ultivated forestland had the relationship with elevation, slope, and the distance of road, river, and previous unlawful cultivated forestland. However, it was strongly correlated with the distance of road and previous unlawful cultivated forestland, and less correlated with elevation, slope, and the distance of river. Finally, the spatial simulation of unlawful cultivated forestlands illustrated that the distribution of unlawful cultivated forestlands was gradually expanded, and the disturbance level was differences at the significance level. From the result, obviously the integration of Markov chain and binomial logit regression analysis is a feasible approach to predict the overall landscape change in future and to simulate the spatial distribution of unlawful cultivated forestland. Also, the Shannon t-test is a good way to monitor and assess the disturbance effect of unlawful cultivated forestland on future landscape. The above result can be provided to Lien-Hwa-Chih Experimental Forest for the reference of forest management in order to prevent the occurrence of unlawful cultivated forestland.

Heat island effect is a regional thermal anomaly that typically exhibits higher air temperature in urban areas than in their surrounding areas. The anomaly can reach 3 - 6oC in summer. Its occurrence is not only to produce uncomfortably hot air, but also to increase the demand of air conditioning and to degrade the air quality. In this study, we use LANDSAT multi-spectral data to estimate the surface parameters and combine the S-SEBI algorithm with sensible heat flux equation to derive the air temperature (1.5 m), which is then compared with in situ measured air temperature from Central Weather Bureau (CWB). The correlation between the estimated and measured temperatures is 0.86. Finally, we use the air temperature data to quantify the magnitude of the urban heat island effect. The magnitude varies about from 0.6 oC to 3.2oC. As expected, higher temperatures are found around Taichung city, Taichung industrial zone, and Taichung harbor.

The protection forests of Taipei City, established in 1907 with a total area of 2,349 ha, consists of three categories: water resource conservation, soil protection, and scenic area protection. The social and natural environment had changed drastically in the past few decades, however, the maps and cadastral records of these forests were not updated accordingly. Therefore, a thorough investigation is needed to understand the conditions of these protection forests. In this study, digital mapping techniques were used to survey the boundaries, status, and land use of the protection forests in order to produce protection forest base maps. Moreover, a protection forest management system was developed using GIS software ArcView and ArcIMS. The results indicate that GPS and total stations are very useful for obtaining high-quality data, and topographic maps, cadastre, aerial photos, and digital orthoimages can be used for mapping land use types. Being capable of managing, presenting, and analyzing pertinent data to provide useful information promptly, GIS is a good tool for decision making for protection forests management.