Object recognition, as one of the most fundamental and challenging problems in high-resolution remote sensing image interpretation, has received increasing attention in recent years. However, most conventional object recognition pipelines aim to recognize instances with bounding boxes in a supervised learning strategy, which require intensive and manual labor for instance annotation creation. In this paper, we propose a weakly supervised learning method to alleviate this problem. The core idea of our method is to recognize multiple objects in an image using only image-level semantic labels and indicate the recognized objects with location points instead of box extent. Specifically, a deep convolutional neural network is first trained to perform semantic scene classification, of which the result is employed for the categorical determination of objects in an image. Then, by back-propagating the categorical feature from the fully connected layer to the deep convolutional layer, the categorical and spatial information of an image are combined to obtain an object discriminative localization map, which can effectively indicate the salient regions of objects. Next, a dynamic updating method of local response extremum is proposed to further determine the locations of objects in an image. Finally, extensive experiments are conducted to localize aircraft and oiltanks in remote sensing images based on different convolutional neural networks. Experimental results show that the proposed method outperforms the-state-of-the-art methods, achieving the precision, recall, and $F_1$-score at 94.50%, 88.79%, and 91.56% for aircraft localization and 89.12%, 83.04%, and 85.97% for oiltank localization, respectively. We hope that our work could serve as a basic reference for remote sensing object localization via a weakly supervised strategy and provide new opportunities for further research.

Against the backdrop of the continuing spread of cross-border telecommunications fraud and the scarcity of conventional data, this study examines how publicly accessible multi-source data can be used to monitor and characterize the spatiotemporal evolution of known or suspected fraud compounds under data-scarce conditions. Focusing on eight representative and manually delineated fraud compounds in Myanmar’s border regions, it develops a three-dimensional “space–time–representation” framework that integrates 2017–2025 VIIRS nighttime lights (NTL), Sentinel-1/2-derived built-up area (BUA) expansion, and Baidu Index time series as a proxy for Chinese-language public attention. Dynamic time warping (DTW) is used to examine temporal alignment between NTL dynamics and public attention, while spatial autocorrelation and exploratory grouping based on DTW dissimilarity are applied to characterize spatial clustering and heterogeneous temporal patterns. The results show that BUA expansion and nighttime brightness are not uniformly coupled across compounds and years. After excluding partial-year 2025 observations, the complete-year 2017–2024 park–year analysis indicates weak aggregate associations between BUA and annual mean NTL, highlighting the need to interpret the two indicators as complementary but non-equivalent proxies. NTL changes do not necessarily evolve synchronously with changes in public attention, suggesting potential value for monitoring and prioritization rather than direct early-warning or verification of illicit activities. The global Moran’s I of 0.2476 indicates significant spatial clustering of nighttime-light values within the defined compound–buffer analysis units. DTW alignment reveals stronger consistency between nighttime lights and Baidu Index attention in the Daxiang and Dingsheng compounds, whereas KK Park and Yatai City show weaker public-search responses relative to observed NTL dynamics. Exploratory grouping identifies three patterns of NTL–public attention correspondence, highlighting heterogeneous relationships between observed nighttime-light dynamics and external information visibility. The study demonstrates the complementarity of remote sensing and social sensing data under data scarcity and proposes a replicable approach for remote monitoring, typological characterization, and evidence-informed governance prioritization of known or suspected fraud compounds. The proposed approach should be interpreted as a screening and analytical framework that provides directional evidence rather than as direct verification of illicit activity or compound-specific operational status.