Neural Radiance Fields (NeRF) use multi-view images for 3D scene representation, demonstrating remarkable performance. As one of the primary sources of multi-view images, multi-camera systems encounter challenges such...
详细信息
Object detection and pose estimation are difficult tasks in robotics and autonomous driving. Existing object detection and pose estimation methods mostly adopt the same-dimensional data for training. For example, 2D o...
详细信息
Fisheye cameras suffer from image distortion while having a large field of view(LFOV). And this fact leads to poor performance on some fisheye vision tasks. One of the solutions is to optimize the current vision algor...
详细信息
Modern object detectors take advantage of rectangular bounding boxes as a conventional way to represent objects. When it comes to fisheye images, rectangular boxes involve more background noise rather than semantic in...
详细信息
In the context of unstructured and unknown environment, the autonomous navigation still faces many challenges, such as assessing rough terrain and deciding how to safely navigate complex terrain. In this work, we prop...
详细信息
ISBN:
(数字)9798350377705
ISBN:
(纸本)9798350377712
In the context of unstructured and unknown environment, the autonomous navigation still faces many challenges, such as assessing rough terrain and deciding how to safely navigate complex terrain. In this work, we propose a robust and practical off-road navigation framework that has been successfully deployed on a vibroseis truck for land exploration. First, in degraded wild scenes, a tightly coupled lidar-GNSS-inertial fusion odometry and mapping framework is adopted to construct a local point cloud map around the vehicle in real-time and provide precise localization. Then, based on amplitude-frequency characteristic analysis and point cloud PCA, a multi-layer terrain assessment map containing terrain roughness, obstacles and slope information is obtained. Finally, combining Gaussian distribution based adaptive sampler and Bayesian sequentially updated proposal distribution, a local graph is efficiently built to obtain multiple path solutions under constrained conditions. Both simulations and field experiments show that the proposed navigation framework can decide how to travel on a flat road even in harsh terrain conditions, naturally suppressing frequent attitude angle changes and preventing vehicle accidents.
Target Detection is one of the most important tasks in Computer Vision, which has broad application prospects in many scenes. For the past few years, great progress has been made in this field with the rapid developme...
详细信息
ISBN:
(纸本)9781665478977
Target Detection is one of the most important tasks in Computer Vision, which has broad application prospects in many scenes. For the past few years, great progress has been made in this field with the rapid development of deep learning technology. However, it’s still a big challenge at complex scenarios such as dim environment for traditional single-modal visible images. To address this problem, researchers introduce thermal images as additional modal in consideration of that thermal cameras are less susceptible to interference and explore how to fuse the two modalities information effectively. Nevertheless, the lack of large labeled and high-quality visible-thermal datasets hampers the usage of convolutional neural networks for detection. Therefore, we propose to use image-to-image translation model combined with a differentiable data augmentation method to generate fake thermal images from labeled visible images and use multi-modal target detection model to prove the validity of the method. Our experiment results show that our method can provide us a large labeled dataset of synthetic visible-thermal image pairs with better generalization and the introduction of thermal modality can obtain a better performance than single modality.
Continuous trajectory tracking control of quadrotors is complicated when considering noise from the environment. Due to the difficulty in modeling the environmental dynamics, tracking methodologies based on convention...
详细信息
A wheel-legged robot is equipped with Stewart parallel mechanism, constituting a reconfigurable robot which can change its wheelbase, robot body height, and achieve omnidirectional steering. The legged character effec...
详细信息
ISBN:
(纸本)9798400712647
A wheel-legged robot is equipped with Stewart parallel mechanism, constituting a reconfigurable robot which can change its wheelbase, robot body height, and achieve omnidirectional steering. The legged character effectively improves the terrain adaptability, which concerns our planning concentration. We introduced an optimization-based whole-body trajectory planning algorithm to navigate robot in rugged terrain. The planner combines terrain data and stability, allowing lower-level motion generator and controller to operate more efficiently. The Model Predictive control(MPC)-based method updates the footholds and CoG trajectories, which builds upon the support polygon constraints on optimization. The simulations of methodology working in several structure-obstacle scene demonstrated and compared the availability of approach.
A distributed fault-tolerant formation control law is designed in this paper for multi-agent systems under external disturbances and actuator faults including time-varying loss of effectiveness faults. The initial pos...
详细信息
The study primarily delves into addressing the data filtering challenge inherent in the silicon single crystal growth process by proposing a multi-sensor integrated detection approach. This approach is designed to all...
详细信息
ISBN:
(数字)9798350363173
ISBN:
(纸本)9798350363180
The study primarily delves into addressing the data filtering challenge inherent in the silicon single crystal growth process by proposing a multi-sensor integrated detection approach. This approach is designed to alleviate the diminished filtering accuracy resulting from the reliance on a single sensor and the conventional Kalman filtering algorithm. By utilizing an enhanced adaptive Kalman filtering algorithm and the maximum likelihood estimation criterion to incorporate the innovation variance directly into the gain calculation of the Kalman filter, dynamic adjustment of the estimation model is achieved while interference from system noise and measurement noise is reduced. Finally, simulation-based comparative experiments between the adaptive and conventional Kalman filtering algorithms are performed. The results show that the performance of the adaptive Kalman filtering algorithm is superior to that of the conventional Kalman filtering algorithm in terms of accuracy and stability when detecting signals from multiple sensors. The research findings furnish dependable data support for refining the decision-making process in crystal growth technology, augmenting the precision and scientific integrity of crystal safety production scheduling and decision-making endeavors.
暂无评论