Abstract: Ochotona curzoniae is one of the main creature calamities of meadow biogeocenose in Qinghai-Tibet Plateau.The conventional methods for pika prevention are costly in both material resources and manpower.Moreover they are difficult to practice in long-term and continuous investigations and prevention.In this paper,a plateaus pika identification system based on image processing and pattern recognition is proposed.Additionally,image preconditioning methods like gradation,binarization,edge detection and image smoothing are also discussed in detail.The holistic gradation histogram threshold technique is used for binarization while the Median filtering technique is used for image smoothing.This system has the advantage to achieve dynamic and continuous monitoring of Ochotona curzoniae in close quarters.

Key words: Ochotona curzoniae, Image processing, Monitor

摘要： 高原鼠兔(Ochotona curzoniae)是破坏青藏高原草地生态的主要生物灾害之一。对高原鼠兔传统的监测与鼠情调查方法需要耗费大量的人力物力,难以实现对鼠兔长期、连续的观测研究。为此,提出了一种基于图像处理技术的高原鼠兔自动识别方案,把图像处理技术应用到鼠兔识别系统中,获取对鼠兔监测与调查的数据,为鼠兔的防治提供理论依据。与传统的监测与调查方法相比较,该方案能够实现对鼠类活动无接触、动态、连续的监测,对鼠兔的防治以及改善高原的草地生态环境具有重要意义。

关键词: 高原鼠兔, 图像处理, 监测