arXiv subjects

Man Qi

Publications and source records attributed to Man Qi.

3 recordsLinked to original sources

arXiv2025

JRN-Geo: A Joint Perception Network based on RGB and Normal images for Cross-view Geo-localization

Cross-view geo-localization plays a critical role in Unmanned Aerial Vehicle (UAV) localization and navigation. However, significant challenges arise from the drastic viewpoint differences and appearance variations between images. Existing methods predominantly rely on semantic features from RGB images, often neglecting the importance of spatial structural information in capturing viewpoint-invariant features. To address this issue, we incorporate geometric structural information from normal images and introduce a Joint perception network to integrate RGB and Normal images (JRN-Geo). Our approach utilizes a dual-branch feature extraction framework, leveraging a Difference-Aware Fusion Module (DAFM) and Joint-Constrained Interaction Aggregation (JCIA) strategy to enable deep fusion and joint-constrained semantic and structural information representation. Furthermore, we propose a 3D geographic augmentation technique to generate potential viewpoint variation samples, enhancing the network's ability to learn viewpoint-invariant features. Extensive experiments on the University-1652 and SUES-200 datasets validate the robustness of our method against complex viewpoint ariations, achieving state-of-the-art performance.

cs.CV↗

arXiv2024

Watching Grass Grow: Long-term Visual Navigation and Mission Planning for Autonomous Biodiversity Monitoring

We describe a challenging robotics deployment in a complex ecosystem to monitor a rich plant community. The study site is dominated by dynamic grassland vegetation and is thus visually ambiguous and liable to drastic appearance change over the course of a day and especially through the growing season. This dynamism and complexity in appearance seriously impact the stability of the robotics platform, as localisation is a foundational part of that control loop, and so routes must be carefully taught and retaught until autonomy is robust and repeatable. Our system is demonstrated over a 6-week period monitoring the response of grass species to experimental climate change manipulations. We also discuss the applicability of our pipeline to monitor biodiversity in other complex natural settings.

cs.RO↗

arXiv2019

Coexistence under hierarchical resource exploitation: the role of R*-preemption tradeoff

Resource competition theory predicts coexistence and exclusion patterns based on species R*s, the minimum resource values required for a species to persist. A central assumption of the theory is that all species have equal access to resources. However, many systems are characterized by preemption exploitation, where some species deplete resources before their competitors can access them (e.g., asymmetric light competition, contest competition among animals). We hypothesize that coexistence under preemption requires an R*-preemption tradeoff, i.e., the species with the priority access should have a higher R* (lower efficiency). Thus, we developed an extension of resource competition theory to investigate partial and total preemption (in the latter, the preemptor is unaffected by species with lower preemption rank). We found that an R*-preemption tradeoff is a necessary condition for coexistence in all models. Moreover, under total preemption, the tradeoff alone is sufficient for coexistence. In contrast, under partial preemption, more conditions are needed, which restricts the parameter space of coexistence. Finally, we discussed the implications of our finding for seemingly distinct tradeoffs, which we view as special cases of R*-preemption tradeoff. These tradeoffs include the digger-grazer, the competition-colonization, and tradeoffs related to light competition between trees and understories.

q-bio.PE↗