关键词: 高速影像, 风沙颗粒, 灰度, 表观粒径, 多假设追踪算法

Abstract: As a key issue that many scientists have taken efforts to sovle since 1980's, the non-correspondence between dynamic paremeters of turblent wind and saltation flux at high time resolution is not been understood perfectly. A high-precision algorithm to be developed for tracking massive sand particles from 2D high-speed images is very fundamental to understand this issue in terms of individual saltating particle trajectory formation driven by turbluent wind. It can enhance precision of multiple hypothesis tracking by introduction of new parameters as gray level and apparent grain size of saltating particles if they are discriminative and stable well in the 2D images. However, discrimination and stability of gray level and apparent grain size have not been discussed as so far. As a result, this paper aims to discuss this topic based on 1 000 frames of high-speed images. Gray level of saltating particles varies around 25-255, showing a nice discrimination. Gray level difference of about 80%-85% saltating particles is below one times standard deviation as they impact and rebound, which indicates the parameter sound stable. Range around 77-1 001 μm shows apparent grain size can discriminate one well from anothor particle, while apparent grain size difference of 97%-100% incidence or rebound particles and that of 77% incidence and rebound particles is stable. Only about estimated 15%-23% of saltating particles moves transversely or rotationally based on variations in gray level and apparent grain size, not affecting stability of gray level and apparent grain size of majority of saltating particles. Thus, gray level and apparent grain size can be used as characteristic parameters besides position for the multiple hypothesis tracking.

Key words: high-speed images, aeolian saltation, gray level, apparent grain size, multiple hypothesis tracking