NavBLIP: a visual-language model for enhancing unmanned aerial vehicles navigation and object detection

Frontiers in Neurorobotics, Journal Year: 2025, Volume and Issue: 18

Published: Jan. 24, 2025

Introduction In recent years, Unmanned Aerial Vehicles (UAVs) have increasingly been deployed in various applications such as autonomous navigation, surveillance, and object detection. Traditional methods for UAV navigation detection often relied on either handcrafted features or unimodal deep learning approaches. While these seen some success, they frequently encounter limitations dynamic environments, where robustness computational efficiency become critical real-time performance. Additionally, fail to effectively integrate multimodal inputs, which restricts their adaptability generalization capabilities when facing complex diverse scenarios. Methods To address challenges, we introduce NavBLIP, a novel visual-language model specifically designed enhance by utilizing data. NavBLIP incorporates transfer techniques along with Nuisance-Invariant Multimodal Feature Extraction (NIMFE) module. The NIMFE module plays key role disentangling relevant from intricate visual environmental allowing UAVs swiftly adapt new environments improve accuracy. Furthermore, employs control strategy that dynamically selects context-specific optimize performance, ensuring high-stakes operations. Results discussion Extensive experiments benchmark datasets RefCOCO, CC12M, Openlmages reveal outperforms existing state-of-the-art models terms of accuracy, recall, efficiency. our ablation study emphasizes the significance components boosting model's underscoring NavBLIP's potential are paramount.

Language: Английский

---

Advances in Instance Segmentation: Technologies, Metrics and Applications in Computer Vision

Neurocomputing, Journal Year: 2025, Volume and Issue: unknown, P. 129584 - 129584

Published: Jan. 1, 2025

Language: Английский

---

Review: multi object tracking in livestock - from farm animal management to state-of-the-art methods

animal, Journal Year: 2025, Volume and Issue: unknown, P. 101503 - 101503

Published: April 1, 2025

Language: Английский

---

Image Data Acquisition and Classification of Vannamei Shrimp Cultivation Results Based on Deep Learning

Matrik Jurnal Manajemen Teknik Informatika dan Rekayasa Komputer, Journal Year: 2024, Volume and Issue: 23(3), P. 491 - 508

Published: June 8, 2024

This research aimed to employ deep learning techniques address the classification of Litopenaeus vannamei cultivation results in land ponds and tarpaulin ponds. Despite their similar appearance, distinguishable differences exist various aspects such as color, shape, size, market price between two methods, often leading consumer confusion potential exploitation by irresponsible sellers. To mitigate this challenge, proposed a method utilizing Convolutional Neural Network (CNN) architectures: Visual Geometry Group-16 (VGG-16) Residual Network-50 (ResNet-50), renowned for success image recognition applications. The dataset comprised 2,080 images per class shrimp from both types Augmentation enhanced dataset’s diversity sample reinforcing model’s ability discern morphology variations. Experiments were conducted with rates 0.001 0.0001 on Stochastic Gradient Descent (SGD) Adaptive Moment Estimation (ADAM) optimizers evaluate effectiveness model training. VGG-16 ResNet-50 models trained rate parameter 0.0001, leveraging flexibility reasonable control provided SGD optimizer. Lower values chosen prevent overfitting enhance training stability. evaluation demonstrated promising results, architectures achieving 100% accuracy classifying soil Furthermore, experimental findings highlight superiority using over architectures, underscoring significant impact optimizer selection tasks.

Language: Английский

---

Evaluation of Automated Object-Detection Algorithms for Koala Detection in Infrared Aerial Imagery

Sensors, Journal Year: 2024, Volume and Issue: 24(21), P. 7048 - 7048

Published: Oct. 31, 2024

Effective detection techniques are important for wildlife monitoring and conservation applications especially helpful species that live in complex environments, such as arboreal animals like koalas (

Language: Английский

---

Assessing Mission Reliability for Unmanned Aerial Vehicle System in the Face of Uncertain Shocks

Arabian Journal for Science and Engineering, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 30, 2024

Language: Английский

---