IDBR (Type A): Development of app and web interface for automated anuran recognition and mapping

IDBR（A 型）：开发用于自动无尾星识别和绘图的应用程序和 Web 界面

• 批准号: 1152306
• 负责人: Mark Bush
• 金额: $ 34.87万
• 依托单位: Florida Institute of Technology
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2012
• 资助国家: 美国
• 起止时间: 2012-03-01 至 2018-02-28
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1152306&HistoricalAwards=false
• 关键词: IDBR; Type; Development; app; web

A synergy of disease, climate change, pollution, exotic species and habitat loss, is pressuring anuran populations worldwide. In the USA, while overall anuran population declines are clearly evident, species that will be hardest hit are hard to predict. The spatial and temporal monitoring of anuran populations needs to be improved if the signs of further population failures are to be identified and remedies sought. The ideal means to increase the coverage of data is to put out sensors that would detect frogs, or have highly trained researchers conduct surveys. However, both these options are cost-prohibitive. A more cost-effective means to gather these data is to take advantage of the many volunteers interested in the fate of anurans. The limitation to date has been the time-consuming and inherently variable success in training those volunteers to recognize anuran vocalizations. In the proposed research, machine-learning algorithms and statistical analysis of sonograms will be used to provide reproducible, standardized identification of calling anurans in real time. A novel use of mobile devices (e.g. iPods, phones, and tablets) is proposed to identify and map anurans. The vocal recognition software will be developed as application (app) for mobile devices. By producing an efficient and intuitive app, that integrates with an accompanying social network an army of potential volunteers can be involved in mapping the abundance and range of anurans. While this proposal takes advantage of cutting-edge technology in terms of voice recognition, connectivity and social media, its purpose is not to advance any one of those branches of technology. Rather it applies existing technology in a novel way to increase participation in this aspect of citizen science by more than an order of magnitude, and bring the cost of automated voice recognition down by almost three orders of magnitude (from close to $1000 to ~$1). In providing a standardized means to recognize anuran calls, data quality and the reliability of data gathered by volunteers will be improved. General statement of project importance Frogs and toads are excellent indicators of habitat quality, not just of their spawning sites, but also of adjacent uplands. These creatures are experiencing an unprecedented collapse in populations due to habitat loss, pollution, and novel diseases. As the faltering of amphibian populations may presage wider ecological problems, it is important to document where species are most affected. In the proposed research, an application (app) will be developed for smart phones and other mobile devices that will use cutting edge voice-recognition software to identify frogs and toads. Currently the average age for volunteers conducting anuran surveys is about 50. By using modern social media and portable device technology the goal of this proposal is to engage a younger audience. The affordability of the system will make it appropriate for classroom use, and it has the potential to spur studies in ecology, physics, geography, and math, as the students consider the data, sonograms, and maps that the software will produce. Through active dissemination we believe this app could quickly be adopted by thousands, perhaps tens of thousands of users, many of whom would be new contributors to citizen science.

疾病、气候变化、污染、外来物种和栖息地丧失的协同作用正在给全世界的无尾两栖动物种群带来压力。在美国，虽然无尾两栖类的总体数量明显下降，但受影响最严重的物种很难预测。如果要确定进一步种群失败的迹象并寻求补救措施，就需要改进对无尾两栖类种群的时空监测。增加数据覆盖面的理想方法是安装能够探测青蛙的传感器，或者让训练有素的研究人员进行调查。然而，这两种选择都是成本高昂的。收集这些数据的一个更具成本效益的方法是利用许多对无尾两栖动物命运感兴趣的志愿者。到目前为止的局限性一直是在训练这些志愿者认识无尾两栖动物发声耗时和固有的变量成功。在拟议的研究中，机器学习算法和超声波图的统计分析将用于真实的时间提供可重复的，标准化的无尾两栖动物识别。提出了移动的设备（例如iPod、电话和平板电脑）的新用途以识别和映射无尾两栖动物。语音识别软件将作为移动的设备的应用程序（app）开发。通过制作一个高效直观的应用程序，与一个附带的社交网络相集成，一群潜在的志愿者可以参与绘制无尾两栖动物的丰度和范围。虽然该提案利用了语音识别、连接和社交媒体方面的尖端技术，但其目的并不是推进其中任何一个技术分支。相反，它以一种新颖的方式应用现有技术，将公民科学在这方面的参与度提高了一个数量级以上，并将自动语音识别的成本降低了近三个数量级（从接近1000美元降至1美元）。通过提供一种标准化的方法来识别无尾两栖类的叫声，志愿者收集的数据的质量和可靠性将得到提高。 青蛙和蟾蜍是栖息地质量的优秀指标，不仅是它们的产卵地，也是邻近高地的栖息地。由于栖息地丧失、污染和新疾病，这些生物正在经历前所未有的种群崩溃。由于两栖动物数量的减少可能预示着更广泛的生态问题，因此记录哪些物种受影响最大是很重要的。在拟议的研究中，将为智能手机和其他移动的设备开发一个应用程序（app），该应用程序将使用最先进的语音识别软件来识别青蛙和蟾蜍。目前，从事无尾两栖类调查的志愿者的平均年龄约为50岁。通过使用现代社交媒体和便携式设备技术，该提案的目标是吸引年轻观众。该系统的可负担性将使其适合课堂使用，并且它有可能刺激生态学，物理学，地理学和数学的研究，因为学生考虑该软件将产生的数据，超声波图和地图。通过积极的传播，我们相信这个应用程序可以很快被成千上万的用户采用，其中许多人将成为公民科学的新贡献者。