Optimization of linear structure extraction process from magnetic field data using geological modeling

. This article discusses optimizing the process of extracting linear structure from magnetic field data through the modeling of geological bodies. Lineament analysis is an important tool for geophysical fields interpretation, but the results can be ambiguous and dependent on many factors, including the geological features of the area, analysis method and interpreter's opinion. The article reviews the methods of lineament analysis and types of linear structure extraction, as well as conducting modeling of geological objects such as quartz veins and dikes. The results highlight the influence of geological objects on line structures of geophysical data and help choose the best way to interpret geophysical fields.


Introduction
Lineament analysis of potential fields (gravity and magnetic) is currently one of the key tools for geological structures interpretation [1]. This method is used to identify and analysis linear structures, determine their form, direction, and length. But, the results of lineament analysis and fault tracing can be highly ambiguous [2]. Many factors, including geological features of the studied area, must be taken into account during the analysis.
Geological objects play an important role in the accuracy of lineament analysis results based on magnetic field data. Modeling of these objects can be used to study the characteristics of the magnetic field. The aim of this study is to improve the efficiency of lineament analysis based on magnetic field data using petrophysical model based on geology.
In this study, the methodology of lineament analysis, its ambiguity, and types of lineament extraction are examined. In addition, geological body modeling was carried out for quartz veins and dikes. The results allowes identifying the peculiarities of the influence of geological objects on lineament analysis and determining the need to consider them when interpreting magnetic field data.

Lineament analysis
One of the key aspects in assessing the impact of geological features on the accuracy and reliability of lineament analysis results is the identification of lineament types and the establishment of criteria for their identification. Lineaments can be classified based on various characteristics, such as their degree of expression, geometry, etc.
Various methods can be used for lineament extraction, such as analysis of changes in the magnetic field characteristics, detection of correlation breaks between data, distortion of anomaly propagation, and others ( Figure 1). Some lineaments are well identified by automatic methods [3;4]; some require manual analysis [5;6]. Despite the widespread use of lineament analysis for interpretation geophysical data, this method is not without some ambiguity. In particular, the same linear feature can be interpreted differently, depending on the data processing methods used. Different geological features can exhibit similar characteristics on geophysical maps, which makes their precise identification difficult. Additionally, linear features can be false, i.e., not reflect the presence of geological features, which can lead to erroneous interpretation of the data. It is important to consider not only the results of lineament analysis, but also other geological information.

Geological modelling
Geological modeling allows for a more detailed understanding of the structure and relationship of geological bodies with observed potential fields [7]. The resulting models can serve as a basis for mapping geological bodies. In this work, the software ZondGM3d was used -a program for modeling geological bodies in three-dimensional space based on geophysical data [8]. 3D models of geological bodies, such as quartz veins and dikes, were created using available geophysical data. Initially, a block model was created, which is a volumetric array of data consisting of multiple blocks. Then, information about geological objects, their geometric parameters, and properties were gradually added ( Figure 2). The obtained models can serve as a basis for justifying conducting lineament studies. Their analysis can help to study the features of geological structures, which will allow obtaining a more detailed understanding of the relationships between the geological structure and potential fields.
As a result of modeling, the forward problem was solved, and a magnetic field map was obtained. Random noise was added to the calculated magnetic field. Vertical derivative calculated magnetic field and total horizontal gradient calculated magnetic field was calculated and a lineament analysis was performed ( Figure 3). Various methods and techniques, including machine learning methods, are used to cope with ambiguity in lineament analysis. In addition, different technologies and tools are used for automatic line structures tracing. Figure 4 shows the results of automatic lineament extraction from the magnetic field. It can be seen that the lineaments traced by different automatic methods along the synthesized field, correspond to different features of the geological structure. This improves the interpretation of automatic lineament analysis result and can be used to understand the geology of a real area from magnetic field data. Based on the modeling and automatic tracing, the lineaments corresponding to these geological features were identified on the magnetic field map ( Figure 5).

Lineament analysis of filed data
The results obtained on the model are applied to real data, obtained at the area within the Okhotsk-Chukotka volcanic belt. Mineralization is associated with complex zones of submeridional hydrothermal formations. Gold deposits are mainly observed in quartz-adular veins in east part of area, showed on Figure 6.
Three ore bodies with rich gold ore were identified. Quartz-adular veins mainly have near-meridional or northwest strike directions, which were taken into account during the identification of lineaments; one body is confined to the intersection of faults. A ground magnetic survey (100x10 m) was carried out at the area in west part of area, showed on Figure 6. As a result, the map of the anomaly magnetic field was obtained ( Figure 7). An automatic tracing of the magnetic field map was carried out, which allowed for the identification of axes of negative and positive magnetic anomalies, as well as some peculiarities of the magnetic field transformants (Figure 8). However, it should be noted that automatic tracing may contain errors, especially if there are complex field structures. Therefore, visual verification and additional analysis of the obtained results are necessary [10]. As a result, based on the prior geological information, types of lineament extraction, automatic tracing, and the results of modeling on the original magnetic field map, a lineament analysis was performed, and presumed areas of quartz-vein zones were identified (Figure 9).

Conclusions
The study involved modeling of geological objects such as quartz veins and dikes based on available geological data. Subsequently, lineaments corresponding to these geological objects were identified on the magnetic field map through the modeling. Based on the obtained results, presumptive zones corresponding to quartz veins and dikes were delineated on the original magnetic field map.
Overall, the ambiguity in lineament analysis can be caused not only by the complexity of interpreting linear features but also by limited access to prior data and insufficient measurement accuracy. It is important to consider the uncertainty in interpreting the results of lineament analysis and to use additional research methods.