In order to provide a simple and surreal burgeoning approach of interactive sand painting to ordinary users and artists, this paper presented an interactive algorithm to realistically simulate sand painting based on Ki-nect in real-time. The simulation algorithm consists of three subsystems: a preprocessing subsystem, a user inter-action information acquisition subsystem, and a processing and rendering subsystem in real-time. The preproc-essing subsystem includes the collection, analysis and statistics for real sand painting image and the definition for various brush styles. Both of these two operations need to be implemented only once. The acquisition subsystem as the starting point of our simulation algorithm could interactively recognize user action information, such as user gestures, by using Kinect. Based on the action information, the acquisition subsystem could know the func-tion selection of users, and could recognize the painting style and the stroke paths in users’ painting. The process-ing and rendering subsystem adopted height-filed-based sand accumulation algorithm and sandpile collapse algo-rithm to simulate various painting styles, such as hand/fingertip sweeping, sand pouring and sand leakage. Fur-thermore, the height field of sand canvas would be converted back into a RGB image for the final effects render-ing. The experimental results and the user experience feedbacks reveal that the proposed system and algorithm in this paper can generate sand painting artistic creations realistically, effectively, interactively in real-time.

How to combine the human-and-computer interaction and painting art is a hot topic in the fields of computer graphics, virtual reality, film-television advertisements and games. This paper presents an algorithm for interactive simulation of realistic sand painting based on Kinect in real-time, which consists of three sub-systems: the preprocessing system, the user interaction information capture system, and the real-time processing-and-rendering sand painting system. The experimental results reveal that our sand painting simulation is suitable for realistic, interactive and real-time artistic sand painting.

提出一种基于形态特征的月季单株植株建模及大规模园林场景模拟方法.精确采集月季植株各主要器官的真实形态数据;对这些数据特征进行深入分析,获取诸如黄金分割等形态特性;以这些特性为基础建立单株植株主枝与各层分枝间、枝与叶间的拓扑关系,同时模拟茎、叶、花、刺等各器官以生成完整的植株;通过基于角策略的Poisson Disk Tiles分布模型生成大规模园林月季场景.实验结果表明:算法能够真实、有效地生成多种不同形态的单株月季植株并模拟大规模园林月季场景,具有广泛应用价值和广阔发展前景.