Notes on the modeling of debris-flow surface images

Nowadays, debris flows cause a lot of damage all over the world. Surface images of the flow, recorded by a video camera, can provide important information for drawing hazard maps or constructing sediment control dams. We have proposed an Auto Regressive (AR) type image model to build a synthetic surface image of debris flows, and have evaluated various measurement methods using the obtained artificial random image model. In this paper, we first estimate the parameter of the image model by the least square error criterion applied to a real debris flow image. Secondly, we study some characteristics of the surface image of the debris flow such as particle size of the image, isotropy of the image, etc. from the obtained model. Particle size of the debris flow image is estimated correctly from our model in most cases and the surface images of many debris flows are not isotropic.