【摘要】 本文阐述了一种层次化的菌落自动检测方法。该方法以计算机视觉和模式识别理论为依据,通过从宏观到微观的层次化图像分析方法,逐步实现了菌落区域的粗定位和细分割。在此基础上,实现了菌落区域的自动检测和计数。实验证明,该方法可以快速准确的实现菌落的自动计数,从而改进现有基于人力的"平板检测"菌落计数法。更多还原

【Abstract】 The paper proposes a hierarchical automatic detection method of colony.Based on computer vision and pattern recognition,the proposed method can localize the colony region sketchily and segment the colony from background detailedly.Therefore,we can detect and count colony conveniently with this method. Experimental results showed that the proposed method has high accuracy and efficiency and could be used to improve the traditional plate - based colony counting method.更多还原