With the continuous advancement in lacustrine shale oil exploration and development in China, traditional reservoir evaluation methods are facing a series of challenges in microscale characterization. In this study, we analyze the advantages and limitations of existing reservoir evaluation techniques and methods. Accordingly, a more intelligent, comprehensive shale reservoir evaluation method that integrates multidimensional data is proposed. Based on the Adaptive Pyramid Context Network (APCNet) for semantic segmentation, combined with the previously independently developed shale pore-fracture segmentation network (ShaleSeger), this method enables intelligent segmentation of minerals and pore structures within reservoirs. By further integrating image processing techniques with mathematical statistics, the method allows for both the quantitative calculation of the shale brittleness index and the fine-scale characterization of pore structures. The analytical results indicate that this proposed method serves to provide more specific, comprehensive, and quantitative analytical data for shale oil and gas exploration and development, as well as sweet spot identification. These analytical data in turn facilitate the quantitative evaluation of resource potential in shale hydrocarbon reservoirs and assist in the comprehensive assessment of relevant technical difficulties and economic benefits. The systematic solution established based on the proposed method offers a reliable basis for intelligent decision-making in the efficient exploration of lacustrine shale oil and gas. Additionally, this study presents a thorough analysis of challenges associated with current intelligent analysis techniques for lacustrine shale reservoirs and points out focus for future research.

Rock particle segmentation is essential for identifying thin sections accurately, yet tuning deep learning models for this task is challenging. Existing studies have attempted auto-tuning techniques but face three limitations, including limited search space, inefficient search, and high demand for storage space. In this paper, we introduce HyperRPS, a novel multi-fidelity hyper-parameter tuning framework tailored for rock particle segmentation, which addresses these issues simultaneously. First, by augmenting the search space with diverse hyper-parameters, we unleash the potential of the segmentation model. Second, to accelerate the hyper-parameter search, we propose a tuning algorithm leveraging multi-fidelity surrogate fitting and time constraint modeling. Third, we develop a storage management technique to reduce the space requirements for asynchronous scheduling. Experimental results on both public and rock particle segmentation datasets demonstrate HyperRPS’s superior performance. Specifically, on the rock particle segmentation dataset, HyperRPS achieves notable improvements of 8.26% and 2.78% on average precision when tuning Mask2Former (Swin-B) and Mask2Former (Swin-L), respectively, outperforming existing state-of-the-art methods.