Pegmatite dikes are important prospecting indicators for rare-metal deposits, whereas traditional methods for pegmatite dike identification are constrained by the limited capability of human visual interpretation to capture information from remote sensing imagery, resulting in low identification accuracy and efficiency. In recent years, global research on semantic segmentation of different surface features and remote sensing-based mineral exploration using deep learning methods and high-resolution remote sensing imagery has made significant progress; however, studies on surface-exposed geological bodies such as pegmatite dikes remain highly insufficient. To address the key problem of efficiently identifying pegmatite dikes in remote sensing imagery, this study proposes an improved model based on UNet++, termed GAD-UNet++. In the field of remote sensing geology, this study constructed a pegmatite dike semantic segmentation dataset based on high-resolution RGB imagery by using 0.66 m RGB imagery for visual delineation and ZY1F hyperspectral data for spectral constraint and label refinement; on this basis, semantic segmentation of surface pegmatite dikes in the Yilanlike area of the South Tianshan Mountains, Xinjiang, was conducted using RGB remote sensing image patches as model input. Specifically, because pegmatite dikes are small targets characterized by slender structures, indistinct boundaries, and sparse regional distribution, this study introduced a lightweight feature extraction structure (GhostNetV2) and a long-range dependency attention module (DFC) at the encoder stage, and further incorporated the Coordinate Attention module (CA) to enhance spatial localization and boundary representation of the targets. Finally, focal cross-entropy loss and a deep supervision strategy were adopted to improve the accuracy of semantic information extraction for pegmatite dikes, as well as the training stability and segmentation accuracy under class-imbalance conditions. The results show that the proposed model achieved an mIoU of 93.11% and an F1-score of 94.95% on the test set. Compared with existing semantic segmentation models, the proposed model achieved superior performance in both identification accuracy and computational efficiency for pegmatite dikes. In addition, this study delineated 18 potential pegmatite dike enrichment zones in the Yilanlike area, providing technical support for remote sensing-based rare-metal prospecting and geological interpretation in the study area.
The Axi gold deposit, which is located in the Tulasu Basin of the West Tianshan orogenic belt in Northwest China, features vein-type ore bodies hosted in radial structural fractures formed due to volcanic activity. The deposit experienced three distinct mineralization stages: Stage I, characterized by the microcrystalline quartz–pyrite crust; Stage II, characterized by quartz–sulfide–native gold veins; and Stage III, characterized by quartz–carbonate veins. Fluid inclusion studies have identified four types of inclusions: pure vapor, vapor-rich, liquid-rich, and pure liquid. The number of vapor-rich inclusions decreases when moving from Stage I to Stage III, whereas the number of liquid-rich inclusions increases. The fluid temperature gradually decreases from 178–225 °C in Stage I to 151–193 °C in Stage II and further to 123–161 °C in Stage III, whereas the fluid salinity decreases slightly from 2.1%–5.1% wt.% NaCl eqv to 1.4%–4.6% wt.% NaCl eqv and finally to 0.5%–3.7% wt.% NaCl eqv. As suggested by the results of the oxygen, hydrogen, and carbon isotope analyses, the ore-forming fluids were primarily meteoric water. Sulfur isotopic compositions indicate a single deep mantle source. The lead isotopic compositions closely resemble those of Dahalajunshan Formation volcanic rocks, indicating that these rocks were the primary source of the ore-forming material. In addition, gold mineralization formed in a Devonian–Early Carboniferous volcanic arc environment. Element enrichment was mainly caused by the circulation of heated meteoric water through the volcanic strata, while fluid boiling and water–rock interactions were the main mechanisms driving element precipitation. The integrated model developed in this study underscores the intricate interplay between volcanic processes and meteoric fluids during the formation of the Axi gold deposit, offering a robust framework for an understanding of the formation processes and enhancing the predictive exploration models in analogous geological settings.
The Xiaorequanzi Cu deposit is located in the western part of the Dananhu–Tousuquan Island arc in eastern Tianshan, Xinjiang. It includes stratiform and epithermal-related veinlet mineralization. However, the genesis of this deposit remains controversial. Therefore, fluid inclusions, H–O isotopes, in situ S, and trace elements in pyrite were employed in this study to constrain the origins of the deposit. The Xiaorequanzi Cu deposit’s mineralization stages can be categorized into the following three phases: I. volcanogenic massive sulfide (VMS) mineralization; II. quartz–chalcopyrite–pyrite; and III. quartz–chalcopyrite–sphalerite stages. Fluid inclusion studies suggest that Stage I is distinguished by high-temperature (peak: 320–360 °C) and moderate-salinity (peak: 7–9 wt%) fluids belonging to the H2O–NaCl ± CO2 system. Stages II–III only exhibit vapor–liquid inclusions, with mineralizing fluids belonging to the medium-to-low-temperature (Stage II peak: 160–180 °C; Stage III peak: 120–130 °C) and medium-to-low-salinity (Stage II peak: 5–7 wt%; Atage III peak: 4–6 wt%) H2O–NaCl system. The H–O isotopic data suggest that mineralizing fluid in Stage I is a blend of magmatic and paleo-seawater sources, while in Stages II–III, meteoric water predominates, accompanied by low mineralizing temperatures. In situ S isotope results indicate that the source of mineralizing materials in Stage I (2.52–4.48‰) were magmatic rocks, whereas the markedly higher δ34S values in stages II–III (4.68–6.60‰) suggest sulfur isotope leaching from sedimentary rocks by meteoric water as the main source. The LA–ICP–MS data of pyrite in the Xiaorequanzi Cu deposit suggest that Py1 was formed through volcanic processes, whereas Py2 and Py3 exhibited epithermal characteristics. Throughout the mineralization process, a trend in increasing oxygen and decreasing sulfur fugacity occurred, accompanied by a decreased mineralization temperature. This observation corresponds with the temperature data derived from the fluid inclusions. Additionally, the principal components of different generations of pyrite segregated as two clusters representing the VMS (Stage I) and epithermal mineralization (stages II–III). In summary, based on comprehensive research and previous geochronological studies, it is suggested that the Xiaorequanzi Cu deposit experienced two mineralization stages. The early stage is related to the volcanic activity of the Early Carboniferous (354 Ma), whereas the later stage is associated with Carboniferous–Permian (266–264 Ma) volcanic intrusions.
Shaligou, located in the southwestern Altyn Mountains, is a newly discovered lithium-bearing pegmatite deposit that is driving increased exploration in the region. However, the challenging environmental conditions of the Altyn Mountains pose challenges for the exploration of lithium-bearing pegmatites. Consequently, remote sensing technology has emerged as an effective exploration tool. In this study, the spectral data of typical samples were collected to establish a spectral library of rocks and minerals in the study area capable of serving as a foundation for remote sensing analysis. Firstly, ASTER data were utilized successfully for lithological interpretation of the area. Secondly, high-resolution WorldView-3 data with a spatial resolution of 0.31 m were used to establish interpretation criteria for pegmatite dikes. Ground validation results were highly consistent with the remote sensing interpretations, confirming that the use of WorldView-3 data significantly enhances the accuracy of lithium-bearing pegmatite dike identification, providing valuable guidance for further exploration.