|
Design 2023
人工智能绘画技术在街道空间设计中的应用研究
|
Abstract:
目的:利用人工智能绘图技术在街道空间设计应用过程中提升效率,满足高质量效果图渲染需求,以及设计师愉快的设计体验的探索。方法:基于当前热门的开源AI绘画软件“Stable Diffusion”,邀请相关专家学者线下操作该绘图软件的3条绘图路径,图生图、文生图、分割生图(分割生图原理基于图像色块分割技术和图像重建技术)。建立探索AI绘画技术对于街道空间设计应用中的可用性定量研究任务;收集反馈,针对应用中展示效果图层面的清晰性、一致性、可用性,和应用过程体验的效率性、易学性,进行定量的分析测度。结果:现如今的AI绘画技术可以快速地识别和解决问题,大大缩短设计周期和成本,但不同路径的应用效果在空间设计领域也具有一定差异。其中,图生图路径出图效率最快,而分割生图路径的出图效果更适合环境空间设计用户的使用。根据不同技术路径的影响机制进一步提出优化策略。结论:人工智能绘图技术的发展更迭,推动了街道设计中表现方法的更新。人工智能绘图技术对于街道空间设计相关者,不仅促进设计效率,降低设计成本,实现高效表达与良好设计体验等方面具有重要的现实意义。
Objective: To use artificial intelligence drawing technology to improve efficiency in the application process of street space design, meet the needs of high-quality renderings rendering, and explore the pleasant design experience of designers. Method: Based on the popular open source AI painting software “Stable Diffusion”, relevant experts and scholars are invited to operate three drawing paths of the drawing software, drawing, text, and segmentation (the principle of segmentation drawing is based on image color block segmentation technology and image recon-struction technology). Establish the quantitative research task of exploring the availability of AI painting technology for the application of street space design; collect feedback, and make quan-titative analysis and measurement for the clarity, consistency and availability of the rendering map level in the application, and the efficiency and learning of the application process experience. Result: The current AI painting technology can quickly identify and solve problems, greatly shortening the design cycle and cost, but the application effect of different paths also has certain differences in the field of space design. Among them, the drawing path drawing efficiency is the fastest, and the drawing effect of the segmentation drawing path is more suitable for the use of environmental space design users. The optimization strategy is further proposed according to the influence mechanism of different technical pathways. Conclusion: The development and change of artificial intelligence mapping technology has promoted the update of the expression methods in street design. Artificial intelligence drawing technology is of great practical significance for street space design stakeholders, which not only promotes design efficiency, reduces design cost, and realizes efficient expression and good design experience.
[1] | 刘倩. 人工智能技术在影视传媒领域的应用[J]. 现代视听, 2018(11): 25-29. |
[2] | 江剑英. 从街道规划设计到城市规划设计——基于《横琴新区街道设计导则》编制的探讨[C]//中国城市规划学会、重庆市人民政府. 活力城乡 美好人居——2019中国城市规划年会论文集(14规划实施与管理). 中国城市规划学会、重庆市人民政府: 中国城市规划学会, 2019: 260-273. |
[3] | Mohamed, F., Islam, K. and Ahmed, A. (2021) CAD Using Preference Compared to Hand Drafting in Architectural Working Drawings Coursework. Ain Shams Engineering Journal, 12, 3331-3338.
https://doi.org/10.1016/j.asej.2021.01.016 |
[4] | 吴丹子. 基于CAD SketchUp Photoshop三位一体的新农村景观设计实例[J]. 农业网络信息, 2012(9): 8-11. |
[5] | 郑华海, 刘匀, 李元齐. BIM技术研究与应用现状[J]. 结构工程师, 2015, 31(4): 233-241. |
[6] | 覃京燕, 刘碧雨, 张盈盈. 游戏设计中声效应用对于用户交互体验的影响研究[J]. 包装工程, 2011, 32(22): 23-26. |
[7] | 吴振天. 人工智能绘画技术中文字指称问题研究[J]. 美与时代(上), 2020(5): 102-103. |
[8] | 郑亚南. 人工智能对图像文化的解构与重建[J]. 四川戏剧, 2019(8): 34-36. |
[9] | 麻斯亮, 魏福义. 人工智能技术在金融领域的应用: 主要难点与对策建议[J]. 南方金融, 2018(3): 78-84. |
[10] | 郑凯, 王菂. 人工智能在图像生成领域的应用——以Stable Diffusion和ERNIE-ViLG为例[J]. 科技视界, 2022(35): 50-54. |
[11] | 卢经纬, 郭超, 戴星原, 缪青海, 王兴霞, 杨静, 王飞跃. 问答ChatGPT之后: 超大预训练模型的机遇和挑战[J]. 自动化学报, 2023, 49(4): 705-717. |
[12] | 翟尤, 李娟. AIGC发展路径思考: 大模型工具化普及迎来新机遇[J]. 互联网天地, 2022(11): 22-27. |
[13] | 蒋芳. 基于“共融共享”理念的城市街道空间设计[J]. 建筑经济, 2021, 42(3): 133-134. |
[14] | 李彦冬, 郝宗波, 雷航. 卷积神经网络研究综述[J]. 计算机应用, 2016, 36(9): 2508-2515+2565. |
[15] | 王建冬. 国外可用性研究进展述评[J]. 现代图书情报技术, 2009(9): 7-16. |
[16] | Sun, Y., Yang, C.-H., Lyu, Y. and Lin, R. (2022) From Pigments to Pixels: A Comparison of Human and AI Painting. Applied Sciences, 12, 3724. https://doi.org/10.3390/app12083724 |
[17] | 郭炯, 郝建江. 人工智能环境下的学习发生机制[J]. 现代远程教育研究, 2019, 31(5): 32-38. |