【作者】 洪涛；

【导师】 曹翠萍； 郑勇奇；

【作者基本信息】 安徽农业大学 ， 林木遗传育种， 2024， 硕士

【副题名】以栎属为案例

【摘要】 植物分类研究对于揭示生物多样性和维护生态平衡具有重要意义,它为保护自然环境、开展生物资源的可持续利用提供了科学基础。而现行的图像识别技术在面对复杂的野外环境时,识别准确率的不足仍然是其瓶颈,严重影响用户的实际应用体验。本研究应运而生,通过研发一款集成植物分类检索表的野外植物识别APP,旨在克服传统拍照识别技术的局限,并且以壳斗科栎属植物为案例,探讨了一种提升识别准确率的新方法。主要研究结果如下:(1)设计完成基于分类检索表的野外植物识别APP总体框架设计,完成低保真原型图的搭建及其交互设计,绘制APP高保真效果图。首先,通过模块化设计方法优化了APP首页和“我的”页面,提高了用户界面的直观性。其次,开展了低保真和高保真原型设计,确立了用户交互流程和界面布局,重点在于简化用户操作过程,加快植物识别速度。最终,依据设计规范完成了高保真效果图的制作,确保了APP的视觉吸引力和使用便捷性。(2)开发完成基于分类检索表的野外植物识别APP检索识别板块,完成该板块界面设计,完成检索识别功能实现和后端管理系统设计。研究首先概述了检索识别模块的设计思路与目标,针对如何高效整合和处理大量植物性状数据的挑战,采用了先进的人工智能算法和动态性状选择机制。在界面设计与用户体验方面,通过界面设计的核心原则,细化用户体验考量,并具体实现了分类检索界面和“根茎叶花果”板块,以及性状说明弹窗功能,显著提升了APP的操作便捷性和识别效率。此外,后端管理系统的建立不仅保障了性状数据库的持续更新与维护,也实现了用户内容的有效监控与管理。在拍照识别与检索识别的衔接上,通过技术创新实现了两者之间的无缝连接,优化了在拍照识别不确定时向检索识别的平滑过渡(3)分析了基于《中国植物志》的植物分类检索表性状数据库的构建过程,拆分并标准化了部分科、壳斗科及栎属的分类检索表,且建立了用于案例检验的数据库,完成了此案例在应用程序中的使用说明编制。通过细致分析壳斗科及其近缘科植物的形态特征,并将这些特征拆分、标准化,能够有效建立系统化的分类框架。本研究设计出了一套集成植物分类检索表的野外植物识别APP,通过以壳斗科栎属植物为案例,展示了该技术在提升识别准确性和用户体验方面的有效性。综上所述,本研究开发了一款集成分类检索与拍照识别技术的野外植物识别APP。基于植物分类检索表开发了检索识别板块。并以壳斗科栎属作为案例初步检验检索识别功能。更多还原

【Abstract】 Plant classification research is crucial for revealing biodiversity and maintaining ecological balance,providing a scientific basis for protecting the natural environment and sustainably utilizing biological resources.However,the current image recognition technology faces a bottleneck in accuracy,especially in complex outdoor environments,severely affecting the practical user experience.This research was initiated to develop a mobile application for wild plant identification integrating plant classification tables,aimed at overcoming the limitations of traditional photo-based identification techniques,using the Quercus genus as a case study to explore a new method for improving identification accuracy.The main research results are as follows:(1)App Framework and Design.The overall framework for the wild plant identification app based on classification tables was designed,including the construction of low-fidelity prototypes and their interaction design,and the creation of high-fidelity visual representations.Initially,the app’s homepage and personal pages were optimized through modular design methods to enhance user interface intuitiveness.Subsequently,both low and high-fidelity prototype designs were developed,establishing user interaction workflows and interface layouts,focusing on simplifying user operations and accelerating plant identification.Finally,high-fidelity visuals were produced according to design standards,ensuring the app’s visual appeal and ease of use.(2)Development of Identification Module.The identification module of the app was completed,covering the design of the module interface,implementation of the identification function,and design of the backend management system.The design rationale and objectives of the identification module were first outlined,addressing the challenges of integrating and processing extensive plant trait data through the adoption of advanced AI algorithms and dynamic trait selection mechanisms.In terms of interface design and user experience,by adhering to core design principles,user experience considerations were refined,and the categorization search interface and the "root,stem,leaf,flower,fruit" sections were implemented,significantly enhancing the app’s usability and identification accuracy.Moreover,the establishment of backend management systems ensures ongoing updates and maintenance of the trait database,as well as effective monitoring and management of user content.A seamless transition between photo identification and search identification was achieved through technological innovation,optimizing the transition when photo identification results are uncertain.(3)Database Analysis and Construction.The process of constructing a trait database based on the "Flora of China" was analyzed,and classification tables for various families,including the Quercus genus,were decomposed and standardized,establishing a database for case validation and completing the usage instructions for this case in the application.A detailed analysis of the morphological characteristics of the Quercus genus and its related families,and the standardization of these traits,effectively established a systematic classification framework.This research designed a wild plant identification app integrating plant classification tables,demonstrating the effectiveness of this technology in enhancing identification accuracy and user experience through the Quercus genus as a case study.In summary,this research developed a wild plant identification app that integrates classification-based search and photographic identification techniques.An identification module based on plant classification tables was developed,and the Quercus genus was used as a case study to preliminarily test the identification functionality.更多还原