多重分形滤波方法和地球化学信息提取技术研究与进展

Progress of multifractal filtering techniques and their applications in geochemical information extraction

圈定地球化学异常是勘查地球化学和环境地球化学研究中的一项基础性工作。笔者通过详细分析近30年来与地球科学密切相关的分形和多重分形等现代非线性理论和新方法的迅速发展和应用,特别是以解决地球化学复杂背景与叠加异常分解难题为例,介绍了广义自相似理论与分形滤波技术的发展和应用,相信在资源、环境、灾害的其他应用领域也将发挥重要影响。文中阐明了多重分形是更具普适性的地球化学元素分布模式,"密度-面积"分形模型是刻画地球化学异常的基础模型,并为多个空间开展分形滤波和分解地球化学复杂背景与叠加异常的理论基础。该方法的显著优势在于将地球化学场的各向异性、尺度不变性、广义自相似性等特征集于一体,具有压制变化性背景干扰,突出局部异常的能力,其中C-A和S-A业已广泛应用于矿产勘查靶区圈定和环境污染模式识别,形成了分解地球化学复杂背景和叠加异常的常用技术,在国内外多个成矿区带取得了显著的找矿效果。

Delineation of geochemical anomalies is essential both in exploration geochemistry and environmental geochemistry. Since the first attempt of using the concentration-area （C-A） multifractal method for separating geochemical anomaly from background, it has become an innovative field of exploration and environmental geochemistry and various fractal filtering techniques have been developed for decomposing overlapping geochemical anomalies from complex backgrounds. This paper reviews the state of the art of the research on distribution patterns of geochemical elements and the approaches to identify geochemical anomalies, and demonstrates that the multifraetal model provides a general distribution type for geochemical elements. The ＂concentration-area （CA）＂ fractal model is a fundamental model for characterizing geochemical anomalies related to mineralization or environment pollution impact. It also provides a base for fractal filtering in various domains such as Fourier domain, Eigen domain and Wash domain. The fractal filtering methods developed in these domains have not only the general advantages of traditional statistical methods but also added attraction as they deal with anisotropic scaling and generalized self-similarity. From application point of view these methods can reduce the influence of the variable background and enhance the local anomalies. The methods have been comraonly applied in delineation of target areas for mineral exploration and in recognition of pollution patterns for environmental assessment.