Surface defect detection is a crucial step in the process of automotive wheel production. However, the task possesses challenges due to complex background and a wide range of defect types. In order to detect the defects on the wheel surface accurately and quickly, this paper proposes a YOLOv5-based algorithm for automotive wheel surface defect detection. The algorithm trains and tests the YOLOv5s model using the self-created automotive wheel surface defect dataset, which contains four kinds of defects: linear, dotted, sludge, pinhole. The extensive experimental results demonstrate that the deep learning network trained by our method can achieve an average accuracy of 71.7% and 57.14 FPS. Our findings prove that this detection algorithm performs better than other common target detection algorithms and meets the real-time requirements of industrial applications.

The deformation of underground gateroads tends to be asymmetric and complex. Traditional instrumentation fails to accurately and conveniently monitor the full cross-sectional deformation of underground gateroads. Here, a full cross-sectional laser scanner was developed together with a visualization software package. The developed system used polar coordinate measuring method and the full cross-sectional measurement was realized by 360° rotation of laser sensor driven by an electrical motor. Later on, the potential impact of gateroad wall flatness, roughness and geometrical profile as well as coal dust environment on the performance of the developed laser scanner were evaluated. The studies show that a high-level flatness is favorable in application of the developed full cross-sectional deformation monitoring system. For a smooth surface of gateroad, the sensor cannot receive reflected light when the incidence angle of laser beam is large, causing data loss. Conversely, the roughness surface shows its priority as the diffuse reflection light can be received by the sensor. With regards to the coal dust in measurement environment, the fine particles of floating coal dust in the air can lead to the loss of measurement data to some certain due to scattering of laser beam.

Biochar has high potential usage in retaining various contaminants, wastewater treatment, and water purification. In this study, three rice husk derived biochars with pyrolysis temperature 300, 400 and 500 ºC, respectively, and pristine rice rusk were used to remove cadmium from aqueous solution. The results showed that about 70% or more of Cd2+ adsorption occurred in the first 960 mins of adsorption kinetics. The Cd2+ adsorption capacity under equilibrium increased with increasing pyrolysis temperature, probably attributed to the increased specific surface area (SSA) under higher pyrolysis temperature noting that significant linear correlation occurred between Cd2+ adsorption capacity and SSA. The Cd2+ adsorption could be best fitted by pseudo-second order model relative to Elovich model and pseudo-first order model. The Cd2+ adsorption rates were higher in film diffusion stage, indicating that film diffusion stage was significant and fast in the early stage of Cd2+ adsorption. In contrast, Cd2+ adsorption by intra-particle diffusion accounted for 47.0%, 47.9% and 43.9% on average of the total Cd2+ adsorption, respectively, indicating that intra-particle diffusion of Cd2+ played a more predominant role in limiting Cd2+ adsorption rate. When reaching Cd2+ desorption equilibrium, removal ratio (RR) values were averaged 0.96, 0.91, and 0.90 under three initial concentrations. More than 90 percentage on average of Cd2+ was removed from aqueous solution by biochars and rice rusk as well, thus biochars can be used to efficiently remove contaminants from aqueous environment. Cation exchange, electrostatic attraction, and the complexation with surface functional groups could be the main dominant mechanisms for Cd2+ adsorption-desorption on biochars.

In this work, an evanescent Bragg grating sensor inscribed in a few-mode planar polymer waveguide was integrated in microchannel structures and characterized with various chemical applications. The planar waveguide and the microchannels consisted of epoxide-based polymers. The Bragg grating structure was post-processed using point-by-point direct inscription technology. By monitoring the central wavelength shift of the reflected Bragg signal, the sensor showed a temperature sensitivity of -47.75 pm/K. Moreover, the functionality of the evanescent field-based measurements is demonstrated with two application examples: refractive index sensing of different aqueous solutions and gas-phase hydrogen concentration detection. For the latter application, the sensor was additionally coated with a functional layer based on palladium nanoparticles. During the refractive index sensing measurement, the sensor achieved a sensitivity of 6.5 nm/RIU from air to 99.9% pure isopropyl alcohol. For the gas-phase hydrogen detection, the coated sensor achieved a reproducible concentration detection up to 4 vol% hydrogen. According to the reported experimental results, the integrated Bragg grating-based waveguide sensor demonstrates high potential for applications based on lab-on-a-chip concept.

Face swapping is an intriguing and intricate task in the field of computer vision. Currently, most mainstream face swapping methods employ face recognition models to extract identity features and inject them into the generation process. Nonetheless, such methods often struggle to effectively transfer identity information, result in generated results failing to achieve a high identity similarity with the source face. Furthermore, if we can accurately disentangle identity information, we can achieve controllable face swapping, thereby providing more choices to users. In pursuit of this goal, we propose a new face swapping framework (ControlFace) based on the disentanglement of identity information. We disentangle the structure and texture of the source face, encoding and characterizing them in the form of feature embeddings separately. According to the semantic level of each feature representation, we inject them into the corresponding feature mapper and fuse them adequately in the latent space of StyleGAN. Owing to such disentanglement of structure and texture, we are able to controllablely transfer parts of the identity features. Extensive experiments and comparisons with state-of-the-art face swapping methods demonstrate the superiority of our face swapping framework in terms of transferring identity information, producing high-quality face images and controllable face swapping.

Soft materials bearing rigid, lightweight, and vibration-dampening properties offer distinct advantages over traditional wooden and metal-based fillings for spent fuel transport casks, due to their low density, tunable structure, excellent mechanical properties, and ease of processing. In this study, a novel type of rigid polyurethane foams is prepared using a conventional polycondensation reaction between isocyanate and hydroxy groups. Moreover, the density and size of the pore in these foams are precisely controlled through simultaneous gas generation. The as-prepared polyurethane exhibits high thermal stability exceeding 185 °C. Lifetime predictions based on thermal testing indicate that these polyurethane foams could last up to over 60 years, which is double lifetime of conventional materials of about 30 years. Due to their occlusive structure, the mechanical properties of these polymeric materials meet the design standards for spent fuel transport casks, with maximum compression and tensile stresses of 6.89 and 1.37 MPa, respectively, at a testing temperature of -40 °C. In addition, these polymers exhibit effective flame retardancy; combustion ceased within 2 s after removal of the ignition source. All in all, this study provides a simple strategy for preparing rigid polymeric foams, presenting them as promising prospects for application in spent fuel transport casks.

Predicting the outcome of a future game between two National Basketball Association (NBA) teams poses a challenging problem of interest to statistical scientists as well as the general public. In this article, we formalize the problem of predicting the game results as a classification problem and apply the principle of maximum entropy to construct NBA maximum entropy (NBAME) model that fits to discrete statistics for NBA games, and then predict the outcomes of NBA playoffs by the NBAME model. The best NBAME model is able to correctly predict the winning team 74.4 percent of the time as compared to some other machine learning algorithms which is correct 69.3 percent of the time.