SBIR Phase I: COWculator: Automated Cattle Counting and Bovine Temperature Screening from Aerial Feedlot Images

这个小型企业创新研究(SBIR)第一阶段项目的更广泛的影响/商业潜力将来自开发一种快速、容易和准确的方法，通过无人机(UAV)在饲养场和牧场清点牛和检测牛病。目前的牛只计数方法非常耗时或不准确，有时两者兼而有之。此外，牛疾病的诊断往往为时已晚，导致饲养场50%的牛死亡，给养牛业造成了19亿美元的经济损失。拟议的技术将利用航空图像来(A)准确和高效地清点牛，(B)在临床症状出现前长达一周的时间内识别生病的奶牛，而不需要在每头牛上安装昂贵的健康监测设备。最终，这项拟议中的技术有望通过自动计算航空图像来更广泛地影响追踪野生动物和濒危物种的方式。这个小型企业创新研究(SBIR)第一阶段项目建议为饲养场会计、营养学家和审核员开发一种基于图像的解决方案来监控牛。该项目将利用饲养场围栏的航拍照片，使用深度学习和传统图像处理工具的组合，自动清点所有牛品种-无论季节和地面条件。此外，该项目将利用航空热像仪测量牛的温度；将开发机器学习工具，以区分与疾病相关的体温升高和与正常混杂因素相关的体温升高。这个第一阶段项目的目标是开发和充分验证牛在饲养场计数的技术，并建立利用航空热像成像来预测牛健康的技术可行性。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project will result from the development of a quick, easy, and accurate way to count cattle and detect bovine illnesses on feedlots and ranches via Unmanned Aerial Vehicles (UAVs). Current methods for counting cattle are extremely time-consuming or inaccurate, and sometimes both. Additionally, bovine illnesses are often diagnosed too late, leading to 50% of cattle mortalities on feedlots and yielding a $1.9 billion economic loss to the cattle industry. The proposed technology will leverage aerial images to (a) count cattle accurately and efficiently and (b) identify ill cows up to one week before clinical symptoms appear without the need to install expensive health-monitoring equipment on each cow. Ultimately, the proposed technology promises to more broadly impact the way wildlife and endangered species are tracked by automating wildlife counting on aerial images. This Small Business Innovation Research (SBIR) Phase I project proposes to develop an imaging-based solution for feedlot accountants, nutritionists, and auditors to monitor cattle. The project will leverage aerial photos of feedlot pens to automatically count all cattle breeds - regardless of season and ground conditions - using a combination of deep learning and traditional image processing tools. Additionally, this project will leverage aerial thermography to measure bovine temperatures; machine learning tools will be developed to differentiate between elevated body temperatures associated with illness and those associated with normal confounding factors. The goals of this Phase I project are to develop and fully validate the technology for cattle counting on feedlots and to establish the technical feasibility of leveraging aerial thermographic imaging for prediction of cattle health.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.