2026-10-10T01:09:59Z https://api.figshare.com/v2/oai

oai:figshare.com:article/33998042 2026-09-25T17:26:56Z category_15 category_21 category_272 category_734 category_69 portal_5 item_type_3 month_year_09_2026

<p>Multi-object tracking performance on the UAVDT-MOT test split. ↑ : higher is better; ↓ : lower is better.</p> Jianhua Gong (530290) Yifu Wang (5614001) Jun Zhang (48506) Defei Du (25111175) Yang Huang (354741) Neuroscience Biotechnology Environmental Sciences not elsewhere classified Biological Sciences not elsewhere classified Inorganic Chemistry xlink "> continuous unmanned aerial vehicles spatial &# 8211 pruned model runs progressive feature pyramid method attains 44 limited compute available lamp structured pruning irregular equipment geometries improved bytetrack introduces giou regression loss dynamic situational awareness detector augments yolov8s demand aerial coverage aware kalman correction adaptive feature extraction gas field surveillance scale semantic conflict integrated lightweight detection based object detection 1 %, 57 tracking framework tailored geographically dispersed oil gas production facilities gas dataset scale variation lightweight uav based inspection 6 %, yet ground uav ), tensorrt fp16 temporally sparse suppresses cross state cues site imagery shared detail reliable monitoring regulatory compliance paper presents intelligent oil head ), extreme target enhanced head embedded deployment deformable c2f constructed oil confirming practicality automated interpretation asset integrity <p>Multi-object tracking performance on the UAVDT-MOT test split. ↑ : higher is better; ↓ : lower is better.</p> 2026-09-25T17:26:35Z Dataset Dataset 10.1371/journal.pone.0358757.t010 https://figshare.com/articles/dataset/_p_Multi-object_tracking_performance_on_the_UAVDT-MOT_test_split_higher_is_better_lower_is_better_p_/33998042 CC BY 4.0