Large-scale low earth orbit (LEO) satellite constellation projects are increasingly adopting inter-satellite links (ISLs) to enable their autonomous and collaborative operation. Due to the large number of satellite constellation network nodes and their continuous movement in orbit, the network nodes may not remain visible to each other at all times, so the ISL-building choices among satellites are diverse and vary with time. As a result, maintaining a network topology requires onboard planning and management. In this paper, we creatively propose an onboard centralized ISL-building planning scheme with the goal of autonomous topology management. A multi-antenna visibility calculation method that takes the antenna installation angle and the turntable rotation threshold into account is provided for the visibility calculation procedure. Additionally, the link-building planning process is modeled using integer linear programming (ILP); however, to tackle the computational complexity problem of ILP, a link-building planning method based on topology stability optimization is presented. The simulation results show that the proposed onboard centralized ISL-building planning scheme can operate among satellites to successfully realize network status collection, visibility calculation, link-building planning, and planning result distribution, as conducted by the dynamic primary satellite. Moreover, the inter-plane link-building planning method based on topology stability optimization improves the network topology stability on the basis of reducing the network delay.

Pixel-level image manipulation localization requires separating weak forensic traces from image content while recovering manipulated regions under post-processing and domain shift. Existing methods do not jointly impose structured discrimination on multiscale region features and edge features drawn from different images. We therefore propose a dual contrastive learning framework built on a Mesorch-based convolutional–Transformer detector. Multiscale Region Contrastive Learning (MRCL) applies intra-image directional contrast to authentic and manipulated samples from four hierarchically fused feature scales. Cross-Image Edge Contrastive Learning (CECL) uses Canny-derived authentic edges and mask-derived manipulated boundary bands to construct within-image contrast, while separate historical queues provide manipulated positives and authentic negatives for cross-image contrast. Both objectives regularize training without changing the active prediction path at inference. Under Protocol-MVSS, models are trained on CASIAv2 and evaluated on COVERAGE, Columbia, NIST16, CASIAv1, and IMD2020. Across seeds 42, 43, and 44, the complete model improves mean F1 and IoU over the Mesorch-based binary-cross-entropy baseline on all five test sets. In the descriptive comparison with representative methods, its three-seed mean is highest for both metrics on NIST16 and IMD2020 and for F1 on CASIAv1, with the second-highest IoU on CASIAv1.