月球撞击坑分类检测研究

Classification Detection of Moon Impact Craters

根据嫦娥二号采集的月球CCD影像,基于分类检测的方法,对撞击坑进行识别检测研究。针对简单撞击坑、撞击盆地和复杂撞击坑的不同形态特征和影像特征,采用不同的滤波和检测算子进行识别。结果表明,对直径在1.5 km以上且光照条件较好的简单撞击坑和撞击盆地识别效果较好;对于复杂撞击坑,由于其形态结构相对复杂,造成图像边缘特征不突出,因此识别率相对较低,有待进一步改进。

According to CCD image from the moon taken by the CE-2,based on classification of detection methods,this paper identified and detected impact craters.For simple crater,impact basin and complex different morphological characteristics and imaging characteristics of impact craters,the paper adopt different filter and detection algorithm for identification.The results show that the diameter of over 1.5 km and illumination condition good simple crater and impact basin has good recognition effect.For complex crater,due to its morphological structure is relatively complex,causing the image edge character is not prominent,so the recognition rate is relatively low,needs further improvement.