ALSwave : a new method for filtering airborne laser scanner data and extracting building features in urban areas | |
| Author | Vu Tuong Thuy |
| Call Number | AIT DISS. no. SR-03-03 |
| Subject(s) | Remote sensing--Equipment and supplies Scanning systems Wavelets (Mathematics) Filters and filtration |
| Note | A dissertation submitted in pa1tial fulfillment of the requirements for the degree of Doctor of Engineering |
| Publisher | Asian Institute of Technology |
| Abstract | Airborne laser scanner (ALS), or Light Detection And Ranging (LIDAR) technique is becoming a common operational tool in remote sensing, photogrammetry, survey and mapping. Its application is also growing in various fields. The technique is capable of providing dense and accurate elevation cloud points. The cost to produce the elevation data from ALS is significant less than other traditional acquisition techniques. As a matter of fact, the commercialization of ALS has expanded rapidly. For the application that requires high density and high accuracy elevation data, such as urban planning and management, ALS is the most potential provider. However, the post-processing of ALS data after surveying flight remains as unsolved problem. Post-processing the ALS data is challenging due to the complex distribution of objects on earth's surface and still in development stage. The prime objective of this research is to find out a new solution for this problem, especially for the ALS acquired data over urban area. Essentially, this study addresses the basic question of how to filter and how to extract the features from ALS data. The problem of filtering ALS data has been investigated so far with varieties of algorithms, but they still suffer from different magnitude of drawbacks. This study proposed a new and improved method, based on multi-resolution approach. Wavelet was used to filter the laser scanner data. This enabled the classification of objects based on their size and the efficiency to filter out unwanted information at a specific resolution. The proposed algorithm, therefore, named ALSwave (Airborne Laser Scanner Wavelet) method. ALSwave was tested on the acquired ALS data from two center urban areas of Tokyo, Japan and Stuttgart, Germany. Two different kinds of ALS data were used. The results showed a well-filtered bare earth surface coupled with acceptable computational time. The comparison between the filtered bare ea1th surface by ALSwave and the manually filtered one produced good RMS (root means squared) error of 0.4 m and 1.3 m for Tokyo data and Stuttgart data, respectively. The extraction of the target features from the laser cloud points has become an interesting research topic with the emergence and implementation of several new algorithms operating on high-density laser points. However, the adoption of lowdensity ALS data in extraction faces the significant difficulties due to the lack of structural and textural information. Moreover, the extraction process becomes more challenging in a dense urban area, which contains lots of skyscrapers, interspersed with a myriad of low and small houses along with crowded outdoor human activities. The study focused on the extraction of buildings and generation of 3D building database. To extract the buildings in a dense urban area from the low-density ALS data with the assistance of extractable information from the aerial photographs, this study designed and implemented another new approach, which was integrated with the wavelet-based filtering (ALSwave). The implemented approach was tested in Shinjuku-ku, Tokyo, Japan, and the result was excellently matched with the existing 2D vector data. The successful operation of the two new approaches demonstrate that waveletbased approach has opened a new way to filter the crude laser data and subsequently generate fast and more accurate digital terrain models. It is also proved as an appropriate approach in eliminating the unnecessary features surrounding the buildings and extracting the buildings. While the proposed approach was successfully tested on urban area to fulfill the objective of this research, it is recommended to further improve it in subsequent studies, including the forest application. |
| Year | 2003 |
| Type | Dissertation |
| School | School of Advanced Technologies (SAT) |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Space Technology Application and Research (SR) |
| Chairperson(s) | Yokoyama, Ryuzo;Yamazaki, Fumio |
| Examination Committee(s) | Honda, Kiyoshi;Apisit Eiumnoh |
| Scholarship Donor(s) | Government of Japan |
| Degree | Thesis (Ph.D.) - Asian Institute of Technology, 2003 |