| Summary: | 碩士 === 國立宜蘭大學 === 土木工程學系碩士班 === 96 === The modern way to acquire land use classification of large area is using photogrammetric and remote sensing technique. With the advance of every technique used in the remote sensing, the sensor’s resolution has been upgrade. However, use remote sensed imagery to extract land cover signature of urban area is still difficult. Recently, the fast development of Airborne LiDAR (Light Detection And Ranging) which uses NIR laser to scan, can provide enormous 3-dimensional height accuracy point data of interest area in a short time. It has become one of the fastest way to generate digital terrain model and many researcher has been integrated it with remote sensed data to obtain batter classification results.
This study investigates the classification results of aerial and satellite images which integrated with height data acquired form LiDAR. Three study subjects are included.
(a) Adding intensity data acquired from LiDAR as NIR image into classification and discusses it application.
(b) Adding height information acquired from LiDAR into classification and exam it effect to the classify result.
(c) Examine the part of missed and fault classification by mathematical or visual method. Find it characteristic and make a summary for the after research.
The experiment revealed that adding intensity data acquired from LiDAR as NIR image into classification can improve the overall accuracy. And the addition of height information, compared to cases with only red、green、blue and multi-spectral imagery, improved the overall accuracy by up to 14% and 1% in aerial image and 5%、4% in satellite image. It proved that integrated LiDAR with image data could improve the classification accuracy certainly.
|
|---|
