Teachers and Design Institute, Harbin Vocational College of Science and Technology, Harbin 150300, China.
Harbin University of Science of Technology, Harbin 150000, China.
Comput Intell Neurosci. 2022 May 25;2022:2820558. doi: 10.1155/2022/2820558. eCollection 2022.
Natural landscape simulation is one of the most popular research contents in computer graphics in the field of research simulation system. The natural landscape animation scene can immerse viewers in the scene, and it is widely used in visual simulation systems. Simulating natural scenery on a computer is a powerful method for studying the rules of the scenery's growth process as well as the mystery of life. The simulation of natural scenery is of particular importance and has far-reaching implications. The most important aspect of optimizing natural landscape design is creating a natural landscape animation that users enjoy. This article proposes a natural landscape animation design method with a self-learning function based on this concept. The random forest model (RF) is introduced in this method and applied to the entire animation design process. RF can generate a learning model with user evaluation as the classification result to guide the automatic design of natural landscape animation, resulting in user-satisfying animations. Simultaneously, the RF-based natural landscape animation design can continuously update the learning model based on user needs and is self-learning. The experimental part of this article verifies the effectiveness of the natural landscape animation design proposed in this article by comparing the selection rate of user satisfaction and dissatisfaction scenes, and further demonstrates that the method in this article can improve the natural landscape. The market application value of user satisfaction generated by animation is high.
自然景观模拟是计算机图形学领域研究模拟系统中最受欢迎的研究内容之一。自然景观动画场景可以让观众沉浸在场景中，广泛应用于视觉模拟系统。在计算机上模拟自然风景是研究风景生长过程规律以及生命奥秘的一种强大方法。自然景观的模拟尤为重要，具有深远的意义。优化自然景观设计最重要的方面是创建用户喜欢的自然景观动画。本文基于这一理念提出了一种具有自学习功能的自然景观动画设计方法。该方法引入了随机森林模型（RF），并将其应用于整个动画设计过程。RF 可以生成一个以用户评价为分类结果的学习模型，指导自然景观动画的自动设计，从而产生用户满意的动画。同时，基于 RF 的自然景观动画设计可以根据用户需求不断更新学习模型，实现自我学习。本文的实验部分通过比较用户满意度和不满意场景的选择率验证了本文提出的自然景观动画设计的有效性，进一步证明了本文方法可以提高自然景观动画的用户满意度。具有较高的市场应用价值。
