In this Letter, we propose a fast speckle noise reduction method with only single reconstructed image based on convolutional neural networks. The proposed network has multi-sized kernels that can capture the component effectively from digital holographic images. For robust performance, is trained large noisy dataset object-dependent and wide range of levels. experimental results show fast, robust, outstanding performance approach.

Slag foaming is a key factor in terms of quality and productivity the electric arc furnace (EAF) steelmaking process. Optimal control slag required, but difficult due to absence practical on-line measuring methods broad process variability. In this study, soft sensor model, which correlates influential variables with height, was developed by using machine learning based long short-term memory (LSTM) networks for modeling sequential nonlinear data. The model validated actual dataset...

Digital holography is the best tool for measuring object information data. Parallel phase-shift digital (PPSDH) particularly suitable moving objects. This method generates images of objects, but image quality diminished because they contain abundant speckle noise. In this paper, we propose an interpolation-based algorithm to reduce levels noise in generated by PPSDH, thus enhancing their quality.

Slag foaming directly affects the productivity and quality of steel during electric arc furnace (EAF) process. Therefore, slag height needs to be monitored in real-time. However, direct measurement is difficult achieve because inside EAF consists harsh environments, i.e., high temperature presence gas dust. A stepwise regression model was created using sensor data from EAF. total 272 operational sets process were used develop validate model. This came 140ton DC-EAF Dongkuk Steel Pohang,...

Slip detection is an essential technology for robotic grippers to autonomously grasp unknown objects and can be achieved using a tactile sensor.In this paper, we propose high-performance multilayer-perceptron-based slip algorithm that utilizes only normal force data obtained by frequency selective surface(FSS) sensor arrays.This in three stages study.First, no-slip training are aggregated such the closely resemble those of real world.Second, most suitable means preprocessing raw output...

Abstract In this paper, we propose a robust and accurate estimation method for the distance required digital holography (DH) reconstruction using convolutional neural networks (CNN) in off-axis DH (off-axis DH). This applies adaptive spectral pooling to reflect distance-related optical characteristics minimize information loss during training phase. Simulations experiments have confirmed that proposed is more than search-based or CNN-based methods.

A unique laser-induced fluorescence reader was developed to detect various antibiotics in foods using a sandwich-type immunoassay platform. For the measurement, magnetic nanoparticles (MNPs) were synthesized size of 10-20 nm be used for extraction antibiotic target by separation and dye-doped silica (color balls) immobilized with antibody 70-80 as probe detection. The system contains diode laser irradiating from bottom top long cell pathlength. In particular, this equipped ratiometric...

This paper presents how the control performance of feature-based visual servo system is improved by utilizing redundant features. The effectiveness features evaluated smallest singular value image Jacobian which closely related to accuracy in world coordinate system. Usefulness verified real time experiments. To investigate control, experiments on dual-arm robot with eight joints are carried out. rotation and translation step responses three, four five examined.

In this paper, we describe the design and optimal control strategy of an under-actuated, anthropomorphic robotic finger. A compliance element (silicon rubber skin) series elastic actuators are incorporated in to reduce number actuators, for finger control. Manufacturing is further simplified by basic joint a 3D printing technique. The proposed metacarpophalangeal (MCP) structure allows flexion-extension adduction-abduction motions be controlled easily independently. However, due...

To reduce speckle noise of parallel phase-shift digital holography, new interpolation method is adapted and it reduces in real time. Simulation results show that image has low high SNR.

Abstract In holographic data storage, tilt is one of the critical disturbances. There are two types tilt: tangential and radial. real systems, radial occur simultaneously. Thus, it difficult to measure compensate for tilt. this study, using a quadratic window, which compares intensity certain area, error signal was generated compensated with proposed algorithm. The image obtained satisfied 0.3 dB tolerance.

To enhance image quality of 2-step parallel phase-shift digital holography, reconstruction algorithm for speckle noise reduction is proposed. Simulation results show that the proposed enhances SNR reconstructed image.

This paper shows how the control performance of feature-based visual servo system is improved by utilizing redundant features. The effectiveness features evaluated smallest singular value image Jacobian which closely related to accuracy in world coordinate system. usefulness verified real time experiments. To investigate system, experiments on dual-arm robot with eight joints are carried out. Translation and rotation step response three, four five examined.

Page-based holographic data storage (HDS) is very sensitive to disturbances such as tilt and media positioning because of extremely high density HDS. As a result, reconstructed images are also affected by disturbances, which represented degradation image signal-to-noise ratio (SNR). Therefore, it important detect compensate the In order there have been many researches using shapes image. However, hard obtain analyze complex iterative experiments required. Furthermore, be done for applying...

The control of a mobile-task robot is implemented with the suitable-decoupling dynamic interactions between two local robots, i.e., task and mobile robot. redundant formed by serial connection it suitable for multiple execution without any constraints on workspace. decoupling process can be initiated when motions static forces required each are described explicitly. In this paper, authors demonstrate performance system controlled independent controllers which cooperating to achieve common...

3D depth measurement of object using digital holography is difficult to be realized because the wavelength shorter than object. In this paper, image reconstruction spherical for analyzed convolutional neural network.