高斯—拉普拉斯算子场源边界识别方法  

Laplacian of Gaussian Operator for Edge Detection

作　　者：王明[1] 何辉[1] 王林飞[1] 刘前坤[1] 

机构地区：[1]中国国土资源航空物探遥感中心,北京100083

出　　处：《物探与化探》2015年第B12期137-143,共7页Geophysical and Geochemical Exploration

基　　金：国家高科技研究发展计划项目"航空地球物理遥感综合探测技术与解释系统研究"(2013AA063905);国家高技术研究发展计划("863"计划)项目(2011AA060501)

摘　　要：高斯—拉普拉斯算子是各向同性的二阶微分算子,具有旋转不变性,对重磁场梯度突变敏感,与异常边界走向无关;利用高斯—拉普拉斯算子消除空间尺度小于高斯空间系数的数据强度变化,压制噪声干扰,计算出梯度的零值点位置圈定重磁场的边界。理论模型和实际资料表明高斯—拉普拉斯算子方法具有边界识别效果好、分辨能力高、边界定位准确等特点,对于不同埋深、不同规模、不同产状的多场源叠加的复杂地质构造,可以清晰地勾勒出梯度相对平缓的微弱异常以及相邻叠加相互干扰的局部异常边界位置,为进一步提高重磁位场局部异常精细解释提供丰富的细节信息,对识别地质体边界、划分大地构造单元、确定断裂带和地质构造走向等具有一定意义。Laplacian of Gaussian Operator with rotation invariance and sensitive to gravity and magnetic field gradient mutation is isotropic second-order differential operator,which has nothing to do with geological boundary objects trend. By calculating the zero value of the Laplacian of Gaussian Operator,the authors delineated the boundary in potential field and extracted geological body distribution,characteristics of fracture structure plane distribution and other geological information. The application of numerical experiments to airborne gravity data indicates that the method can effectively determine the gravity and magnetic source boundaries with higher resolution and accurate positioning. Especially for the multi-sources overlapping geological structure with different depths,different scales and different attitudes,the method can clearly detect the boundaries of weak anomalies with relatively smooth gradient and adjacent local anomalies with superimposed interference. It also can provide more details for further improvement of fine interpretation of the gravity and magnetic local anomalies,and has certain significance for detecting the geological boundaries,devising tectonic units,and determining the fracture zone and the trend of geological structure. All this will facilitate the integrated interpretation.

关 键 词：重磁场源 边界识别 高斯-拉普拉斯算子 噪声干扰 水平梯度 

分 类 号：P631[天文地球—地质矿产勘探]