The blind have so many inconveniences in their lives. In recent years, more and more blind sidewalks are designed and built for them. This is good news for the blind. However, people would ...consciously or unconsciously leave some stuff on the blind sidewalk. Therefore, the occupation of blind sidewalk is a serious problem gradually. We decided to develop an application on smartphone to help the blind to avoid troubles referred above. Within the camera on the smartphone, the application detects obstacles on the blind sidewalk based on image processing and warns the blind that there are obstacles in front of him. This paper proposed an approach on detecting the direction of blind sidewalk after the blind leave the blind sidewalk to avoid the obstacles. We used several algorithms of image processing and conduct several experiments to exam this system.
In recent years, convolutional neural networks (CNNs) have dominated the field of computer vision. Compared to traditional methods, these neural network algorithms exhibit strong biomimetic ...performance advantages in complex visual tasks due to their brain-like structure. However, because some necessary neural characteristics are ignored, these neural network algorithms differ greatly from the computational mechanisms of the brain. This paper starts with extracting basic visual features such as motion direction information from the brain and abstracts, generalizes and models a novel artificial visual system (AVS) for detecting object motion direction in color images based on existing relevant neurophysiological knowledge. We propose a mathematical model and quantification mechanism for each component neuron that generates motion direction selectivity in the visual system using dendritic neuron models, spiking neural network concepts and neurophysiological knowledge of retinal direction-selective ganglion cell pathways. The experiment is based on one million instances of object motion under different environments of noise-free, static and dynamic random noise, dynamic salt-and-pepper noise, dynamic Gaussion noise, and dynamic light changing. In comparison with 4 famous CNNs, LeNet-5, EfficientNetB0, ResNet18, and RegNetX-200MF, we test and verify AVS’s effectiveness, efficiency and strong generalization ability as well as other biomimetic performance advantages including high biological rationality, learning-free capability, interpretability and ease-of-use etc. AVS demonstrates that neuroscience still has important implications for guiding and promoting the development of artificial intelligence technology. Furthermore, AVS firstly provides a successful quantitative reference case study for further understanding motion direction selectivity and other primary cortical encoding characteristics in the brain.
•We provide the first quantitative explanation of how mammalian brains achieve motion direction selectivity in a colored environment, which serves as a successful reference case for understanding and offering insights into other perceptual systems encoded in the cortex.•We propose a novel bio-inspired motion direction detection algorithm, called AVS, with high biological plausibility.•Comparing AVS with 4 classical and state-of-the-art convolutional neural networks, LeNet-5, ResNet18, EfficientNetB0, and RegNetX-200MF, we also validate its remarkable performance in terms of effectiveness, efficiency, generalizability, complexity and biological plausibility.•AVS indicates the great potential for neuroscience theories to further promote neural network algorithms and artificial intelligence technologies.
The correct orientation of an ultrasound (US) probe is one of the main parameters governing the US image quality. With the rise of robotic ultrasound systems (RUSS), methods that can automatically ...compute the orientation promise repeatable, automatic acquisition from predefined angles resulting in high-quality US imaging. In this article, we propose a method to automatically position a US probe orthogonally to the tissue surface, thereby improving sound propagation and enabling RUSS to reach predefined orientations relatively to the surface normal at the contact point. The method relies on the derivation of the underlying mechanical model. Two rotations around orthogonal axes are carried out, while the contact force is being recorded. Then, the force data are fed into the model to estimate the normal direction. Accordingly, the probe orientation can be computed without requiring visual features. The method is applicable to the convex and linear probes. It has been evaluated on a phantom with varying tilt angles and on multiple human tissues (forearm, upper arm, lower back, and leg). As a result, it has outperformed existing methods in terms of accuracy. The mean (<inline-formula><tex-math notation="LaTeX">\pm</tex-math></inline-formula>SD) absolute angular difference on the in-vivo tissues averaged over all anatomies and probe types is <inline-formula><tex-math notation="LaTeX">2.9\pm 1.6^{\circ }</tex-math></inline-formula> and <inline-formula><tex-math notation="LaTeX">2.2\pm 1.5^{\circ }</tex-math></inline-formula> on the phantom.
In this article, we present an inertial switch with three threshold levels, which can provide quantitative acceleration measurements and detect the acceleration direction in the x-y plane. The ...designed device has four movable electrodes attached to the proof mass (one at every side of the square proof mass) and 12 flexible stationary electrodes (three on each side). When the device is subjected to an acceleration input, the movable electrode can contact one or more of the 12 stationary electrodes based on the acceleration magnitude and direction. The acceleration direction can be determined by identifying the individual electrical switches that are activated. The designed switch is simulated using a finite-element model under different acceleration signals of various magnitudes and directions. A device prototype has been fabricated using the SOIMUMPs process and has been tested by a drop-table system under various shock accelerations in different directions. The experimental and simulation results show good agreement indicating that the acceleration direction detection accuracy and resolution improve with the increase in the number of used electrical switches.
Flexible pressure sensing is required for the excellent sensing performance and dexterous manipulation of the measured objects in their potential applications. Particularly, the ability to measure ...and discriminate the direction of force, contact surface, and contact location in real time is crucial for robotics with tactile feedback. Herein, a three-dimensional elastic porous carbon nanotube (CNT) sponge is synthesized by chemical vapor deposition, which is successfully applied in the piezoresistive sensor. In situ scanning electron microscopy study intuitively illustrates the characteristics that the microfibers of the CNT sponge distort and contact with each other under an external force. As a result, new conductive paths are created at the contact points between the CNT microfibers, which provides a basic sensing principle for a piezoresistive sensor. The CNT sponge-based sensor has an ultrahigh sensitivity in a wide pressure range (0–4 kPa for 4015.8 kPa–1), a rapid response time of 120 ms, and excellent durability over 5000 cycles. Moreover, a finlike flexible double-sided electronic skin (e-skin) is fabricated by a simple method to achieve force direction detection, which has potential applications in intelligent wearable devices and human–machine interaction.
Fiber optic surface plasmon resonance (SPR) sensors have achieved a large number of solutions for sensing physical quantities such as refractive index, micro displacement, curvature, strain, etc. ...However, there are few schemes for sensing torsional physical quantities. This article proposes an optical fiber SPR torsion sensor based on a threaded structure. A threaded structure that is prone to deformation during torsion is etched on the optical fiber. After the threaded area, an SPR sensing area is created. Clockwise rotation causes compression deformation in the threaded area, while counterclockwise rotation causes tensile deformation in the threaded area, resulting in a change in the optical transmission mode of the threaded area, changing the SPR incidence angle, causing the corresponding SPR resonance valley to move. Torsion sensing is achieved through the shift of the SPR resonance wavelength, and the torsion direction is determined by the shift direction of wavelength. The experiment indicates that the sensing probe has a clockwise twist sensitivity of 0.9088nm/(rad/m) and a counterclockwise twist sensitivity of -0.8291nm/(rad/m) within the detection range of -47.10rad/nm to 47.10rad/nm, providing a new scheme for directional detection of torsional physical quantities.
•The proposed scheme can detect the distances and directions of multiple targets simultaneously by using specific spectrum symmetry and polarization demultiplexing.•A large direction range of 166.78° ...is realized without direction ambiguity.•When the distance difference or direction difference between the targets is less than the resolution, the targets can still be effectively distinguished.
A photonics-assisted simultaneous detection method of distances and directions for multiple targets is proposed based on polarization multiplexing and spectrum symmetry. A linear frequency modulation transmitting signal is generated by photonic frequency doubling. The echo signals are received by a three-antenna based linear array. The intermediate antenna supplies overlapped reference spectra by orthogonal modulation. According to specific spectrum symmetry, the de-chirped frequencies of echo signals from each target are clearly distinguished through polarization demultiplexing. Thus, distances and directions can be decoupled for multi-targets without direction ambiguity. Three-target proof-of-concept experiments have verified that the average measurement errors of distance and direction are 3.7 cm and 0.39°. The direction measurement range is up to 166.78°. A multi-target detection scheme with a large direction range is provided for the complex electromagnetic environment.
It is challenging to extract satisfactory building outlines from LiDAR data due to the unorganized point cloud and complex building shapes. To solve the issues, a method using adaptive tracing alpha ...shapes (ATAS) and contextual topological optimization is proposed. First, the ATAS method is used to extract sequential boundary points. After that, a method based on point cloud distribution analysis is developed to obtain building dominant directions and line segments of outlines. Finally, regularized outlines are obtained by adjusting all line segments simultaneously under the framework of global energy optimization that considers the geometric errors and contextual geometric relationships between adjacent line segments. Experimental results verify that the proposed ATAS method can efficiently extract sequential boundary points with a minimum 98.49% correctness. In addition, the extracted outlines are attractive and the minimum values of the RMSE, PoLiS, and RCC metrics of the extracted outlines are 0.48 m, 0.44 m, and 0.31 m, respectively, showing the effectiveness of the proposed method.
•Adaptive tracing alpha shapes method (ATAS) is proposed to efficiently extract sequential boundary points directly from unorganized building point clouds with complex shapes, without pre-processing (e.g., triangulation, gridding).•A method based on point cloud distribution analysis is proposed to detect accurate and reliable building dominant directions, which is beneficial to subsequent outline extraction.•The proposed method can extract smooth and attractive outlines from complex buildings by formulating outline regularization as an optimal labeling problem under the framework of global energy optimization, which balances geometric errors of boundary points and contextual geometric relationships between adjacent line segments.
In this paper, we reconsider the problem of detecting a matrix-valued rank-one signal in unknown Gaussian noise, which was previously addressed for the case of sufficient training data. We relax the ...above assumption to the case of limited training data. We re-derive the corresponding generalized likelihood ratio test (GLRT) and two-step GLRT (2S–GLRT) based on certain unitary transformation on the test data. It is shown that the re-derived detectors can work with low sample support. Moreover, in sample-abundant environments the re-derived GLRT is the same as the previously proposed GLRT and the re-derived 2S–GLRT has better detection performance than the previously proposed 2S–GLRT. Numerical examples are provided to demonstrate the effectiveness of the re-derived detectors.
Water wave monitoring is essential in collecting marine parameters for oceanography studies and early warning systems on security and safety, such as drowning detection, weather detection, and gas ...leakage from underwater pipeline detection, because these activities create different water wave patterns that can be further analyzed. The current wave detection methods, such as underwater pressure and resistive sensors, have lower durability as they require direct contact with the water. Monocular camera wave detection can detect the line where water waves propagate. However, a static platform is required to perform monitoring operations. This research aims to develop a continuous capacitive sensor system that a buoy can implement for contactless water surface wave detection and to develop a water wave direction detection algorithm by Principal Component Analysis (PCA) calculation. Capacitive sensors arranged in a circular shape and a compass module are implemented inside the prototype buoy. Each capacitive sensor detects the real-time wave height change by changing the capacitance value. The capacitance values from all the capacitive sensors and the North of the compass sensor are sent to the embedded server for further computations. Processes carried out in the embedded server are initial calibration, centroid calculation, PCA calculations for water wave detection, and data visualization on the webpage. The prototype buoy with a capacitive sensor system and compass sensor developed can detect the four positions tested in the experiment with a mean square error of 38.42° and a mean absolute error of 5.85°.