Abstract The continual management of pest species and their preventative elimination is an ongoing, labour-intensive problem. Bait stations are pivotal in this management process as they are the point of contact between the rodents and the process, since the rodents need to enter the bait stations to consume the poisoned bait. Monitoring of these bait stations provides feedback of the effectiveness of the management process. However, there is a significantly large cost associated with periodically sending pest control experts to check the level of bait in the bait stations. This becomes even more apparent over a large geographical area. In this paper we present a method of reducing the labour component associated with regular bait level monitoring by placing a camera in the bait stations and using machine vision to provide an estimate of the amount of bait remaining and the type of bait in the station. Images of four common bait types were captured under provided artificial light in a closed bait station, and the computer vision algorithms proved effective in identifying the type of bait as well as providing an estimate of the bait level.

中文翻译：

---

啮齿动物诱饵站诱饵水平估计的计算机视觉方法

摘要 有害生物物种的持续管理及其预防性消除是一个持续存在的劳动密集型问题。诱饵站在此管理过程中至关重要，因为它们是啮齿动物和过程之间的接触点，因为啮齿动物需要进入诱饵站以消耗有毒诱饵。对这些诱饵站的监控提供了管理过程有效性的反馈。然而，定期派遣害虫控制专家检查诱饵站中的诱饵水平会产生很大的成本。这在较大的地理区域内变得更加明显。在本文中，我们提出了一种通过在诱饵站放置摄像头并使用机器视觉来估计站内诱饵剩余量和诱饵类型来减少与定期监测饵料水平相关的劳动力的方法。在封闭诱饵站提供的人造光下捕获了四种常见诱饵类型的图像，计算机视觉算法在识别诱饵类型和提供诱饵水平估计方面被证明是有效的。