|
|
|
Abstract
Regressive Morphological Filter for Ground Point Extraction from LiDAR Data
Santhosh Kumar Rajangam
Student
Institute of Remote Sensing, Anna University,
India
Email: annauniv_santhosh@yahoo.com
Harini Sridharan
Student
Institute of Remote Sensing, Anna University
Email: harinis1@gmail.com
Extracting terrain information from raw LiDAR data has been an ongoing research riveting the attention of researches. DTM provides lot of information that is useful for feature extraction. A number of filters have been developed to derive precise terrain models from the raw height data. In this paper, we have developed a regressive morphological filter that uses successively decreasing sizes of structural element to determine lowest point on the terrain. Successive opening and closing operations are compared to prevent extreme sinking of terrain due to the presence of pits or valleys. Two different datasets, one mountainous region and another urban scene have been tested to justify the effectiveness of the filter. The non-ground points have been eliminated and the accuracy of the extracted terrain has been determined to validate the filter. The paper also highlights the importance of ground point information for feature extraction. A case study on road extraction has been presented to illustrate this analysis.
|
|
|