Vol.53-6 CONTENTS

Hiroyuki HASEGAWA

This study aims to investigate the characteristics of LIDAR intensity data for executing land cover classification. The quantitative and qualitative analysis was executed with reflection geometry model and land cover classification data. By these analyses, the LIDAR intensity feasibility for classification was discussed. The survey found that intensity is inversely proportional to angle and distance, though their relation did not agree with the theoretical model. The survey also found that intensity correction with distance and angle is not always applicable, the effect of correction is not significant, and consequently raw intensity value usage is justified. We conclude that some land cover and building materials were separable with intensity data. Old asphalt and grass were separable though cement, slate & zinc, brick, and trees were not easy to recognize. Soil, gravel, and grass could be distinguishable each other.

1. Introduction
2. Reflection model
3. Methods
  3.1 On the ground experiments
  3.2 Aerial observations
4. Result and discussion
  4.1 On the ground experiments
    4.1.1 Raw intensity profile
    4.1.2 Distance dependency
    4.1.3 Angle Dependency
    4.1.4 Intensity characteristics of surface class
5. Aerial survey result
  5.1 Reflection model validation
  5.2 Intensity characteristics of surface class
6. Conclusion
References