Research On 3D Seismic Image Object Identification Based On Ant Tracking

In geological exploration, in order to obtain the distribution of oil and gas resources, seismic data interpretation is need, and fault interpretation is a vital content in seismic data interpretation. Fault is due to underground rock rupture due to a certain degree of pressure, and generate relative displacement or movement along the orientation of faults surface in geology. Fault is the primary reason of the formation and distribution of oil field, which is the channel of when gas and oil moving underground. Either in the later stage of explore oil, gas, and other resources or exploiting these resources, explain and understand the size and of the fault in the ground is of great importance, no matter in evaluating yield and production or explore and manage reservoir. In this paper, we based on the study of seismic data, estimate the size and distribution of fault underground.Conventional fault interpretation is using artificial methods to explain faults, this method not merely overwork interpreter and overtime, but also deeply rely on interpreter’s professional knowledge and experience. Therefore, an automatic fault interpretation method is urgently needed. At present, only in foreign commercial software Petrel which is developed by Schlumberger, have implemented the automatic fault interpretation in 3D seismic image, but in domestic are still blank in this field, the specific implementation process of automatic fault interpretation in Petrel are not available also. In this paper, based on methods adopted in Petrel and ant tracking algorithm, Primary things we achieved as follow:1. Developed 3D seismic image automatic fault recognition based on ant tracking algorithm. In this paper, we refer to thoughts in Petrel, pretreated seismic data with fault enhanced attribute firstly to highlight fault and suppress no-fault, then then used ant tracking algorithm track fault in fault enhanced attribute, realized the automatic fault interpretation method.2. For the problems of the result get by Petrel contain a lot of noise, this article firstly from the point of ant tracking algorithm model, introduced artificial seed point and gradient consistency, give different pheromone update strategy to artificial ants with different confidence,make artificial ants as far as possible on the large fault tracking, decrease the happening of tracking on no-fault. Actual test shows that, this method can effectively suppress noise interference. For the problems of the fault surfaces have some rag, cavity, furcation and gaps, we fill fault surfaces by mathematical morphology and using 3D seismic image skeleton thinning algorithm for fault. Actual test shows that, this method can make the fault surface clearer and more integrated.3. In order to evaluate the effect of the method, we use real seismic work area data for testing. Practical tests show that the method can effectively identify fault in seismic image. Compared with Petrel, the fault identified by the method in this paper, large fault are more clearly and apparent, and effectively reduce the noise interference.