Accurate classification and identification of the detected terrain is basis for long-distance patrol mission planetary rover. But measurement based on vision radar subject to conditions such as light changes dust storms. In this paper, under premise not increasing sensor load existing rover, a recognition method vibration proposed. Firstly, time-frequency domain transformation information realized by fast Fourier transform (FFT), characteristic representation given. Secondly, deep neural...

Fast and accurate road damage detection is essential for the automatization of inspection. This paper describes our solution submitted to Global Road Damage Detection Challenge 2020 IEEE International Conference on Big Data, typical in digital images based deep learning. The recently proposed YOLOv4 chosen as baseline network, while effects data augmentation, transfer learning, Optimized Anchors, their combination are evaluated. We propose a novel generation method generative adversarial...

Accurate global localization is crucial for autonomous navigation and planning.To this end, various GPSaided Visual-Inertial Odometry (GPS-VIO) fusion algorithms are proposed in the literature.This paper presents a novel GPS-VIO system that able to significantly benefit from online calibration of rotational extrinsic parameter between GPS reference frame VIO frame.The behind reason observable.This provides proof through nonlinear observability analysis.We also evaluate algorithm extensively...

Deep learning for cell instance segmentation is a significant research direction in biomedical image analysis. The traditional supervised methods rely on pixel-wise annotation of object images to train the models, which often accompanied by time-consuming and labor-intensive. Various modified methods, based weakly or semi-supervised learning, have been proposed recognize regions only using rough annotations positions. However, it still hard achieve fully unsupervised most approaches that...

This paper introduces a novel GPS-aided visual-wheel odometry (GPS-VWO) for ground robots. The state estimation algorithm tightly fuses visual, wheeled encoder and GPS measurements in the way of Multi-State Constraint Kalman Filter (MSCKF). To avoid accumulating calibration errors over time, proposed calculates extrinsic rotation parameter between global coordinate frame VWO reference online as part process. convergence this is guaranteed by observability analysis verified using real-world...

Motion planning and control of mobile robots rely on high-accuracy estimation pose velocity. Many researchers use motion capture system to estimate the robot state, but this is usually expensive can only be used indoor. For reason, paper proposes an artificial landmarks based cost-effective visual inertial odometry where Iterative Extended Kalman Filter (IEKF) acts as back-end optimizer. Since most tags like bar code facing problem decoding information during detection, we redesign tag which...

In robotics, motion capture systems have been widely used to measure the accuracy of localization algorithms. Moreover, this infrastructure can also be for other computer vision tasks, such as evaluation Visual (-Inertial) SLAM dynamic initialization, multi-object tracking, or automatic annotation. Yet, work optimally, these functionalities require having accurate and reliable spatial-temporal calibration parameters between camera global pose sensor. study, we provide two novel solutions...

Autism Spectrum Disorder is a condition characterized by typical brain development leading to impairments in social skills, communication abilities, repetitive behaviors, and sensory processing. There have been many studies combining MRI images with machine learning algorithms achieve objective diagnosis of autism, but the correlation between white matter autism has not fully utilized. To address this gap, we develop computer-aided diagnostic model focusing on regions employing radiomics...

In robotics, motion capture systems have been widely used to measure the accuracy of localization algorithms. Moreover, this infrastructure can also be for other computer vision tasks, such as evaluation Visual (-Inertial) SLAM dynamic initialization, multi-object tracking, or automatic annotation. Yet, work optimally, these functionalities require having accurate and reliable spatial-temporal calibration parameters between camera global pose sensor. study, we provide two novel solutions...

Accurate disturbance estimation is crucial for reliable robotic physical interaction. To estimate environmental interference in a low-cost and sensorless way (without force sensor), variety of tightly-coupled visual inertial external estimators are proposed the literature. However, existing solutions may suffer from relatively low-frequency preintegration. In this paper, novel estimator designed to overcome issue via high-frequency instantaneous accelerometer update.

Providing self-modelling capabilities to robotic systems that could change their dynamics variables during natural operation, like aerial manipulators, can significantly increase model-based algorithm's resilience and performance. Some samples of those techniques are model predictive control or trajectory planning techniques. This paper aims benchmark how classical identification perform when adapted be used at run-time. self-awareness capability will improve transparency giving the robot...

This paper introduces a novel GPS-aided visual-wheel odometry (GPS-VWO) for ground robots. The state estimation algorithm tightly fuses visual, wheeled encoder and GPS measurements in the way of Multi-State Constraint Kalman Filter (MSCKF). To avoid accumulating calibration errors over time, proposed calculates extrinsic rotation parameter between global coordinate frame VWO reference online as part process. convergence this is guaranteed by observability analysis verified using real-world...

Accurate global localization is crucial for autonomous navigation and planning. To this end, various GPS-aided Visual-Inertial Odometry (GPS-VIO) fusion algorithms are proposed in the literature. This paper presents a novel GPS-VIO system that able to significantly benefit from online calibration of rotational extrinsic parameter between GPS reference frame VIO frame. The behind reason observable. provides proof through nonlinear observability analysis. We also evaluate algorithm extensively...