基于分水岭算法的地震属性异常体边缘检测技术

Edge Detection Technology of Abnormal Seismic Attribute Body Based on Watershed Algorithm

从地震资料中识别采空区时,不易察觉波组特征在地震记录上反映弱的采空区。此外,在采空区周边特殊构造及薄层的影响下,来自不同界面的反射波之间相互干扰,使波形畸变,造成解释困难。为了解决上述问题,将基于分水岭算法的图像分割方法引入地震属性分析中,用计算机自动识别地震属性异常区边缘,提高地震资料解释精度,减少人为因素。以山西五家沟某矿区的三维地震相干属性为例,应用基于分水岭算法的边缘检测方法对其进行处理,圈定采空区的边界范围,经矿方验证效果显著。

Some small abnormal seismic attribution area is easily ignored when shape it, such as goaf. In addition,influenced by special structure and thin layers around the goaf, reflected waves from different interfaces are interfered mutually, then the shape of reflected waves is changed, which attaches difficulty to seismic interpretation. To deal with the problem above, image segmentation method based on watershed algorithm was used to analyse the seismic attribution, and the edge of the abnormal area was detected by computer, whose purpose was to improve the interpretation accuracy of seismic data, and reduce the affects of human factors. Taking a mining area of Wu- jiagou area in Shanxi province as an example,edge detection method based on watershed algorithm was used to outline landslide bound- ary of goaf,the result was proved to be satisfactory.