Search Results (10)

To enhance the synchronous detection of the horizontal and vertical positions of the torch in swing arc narrow gap welding, a torch pose detection (TPD) method is proposed. This approach utilizes passive visual sensing to capture images of the arc on the groove sidewall, using advanced image processing methods to extract and fit the arc contour. The coordinates of the arc contour center point and the highest point are determined through the arc contour fitting line. The torch center position is calculated from the average horizontal coordinates of the arc contour centers in adjacent welding images, while the height position is determined from the vertical coordinate of the arc’s highest point. Experimental validation in both variable and constant groove welding conditions demonstrated the TPD method’s accuracy within 0.32 mm for detecting the torch center position. This method eliminates the need to construct the wire centerline, which was a requirement in previous approaches, thereby reducing the impact of wire straightness on detection accuracy. The proposed TPD method successfully achieves simultaneous detection of the torch center and height positions, laying the foundation for intelligent detection and adaptive control in swing arc narrow gap welding. Full article

The global population’s rapid growth necessitates a 70% increase in agricultural production, posing challenges exacerbated by weed infestation and herbicide drawbacks. To address this, machine learning (ML) models, particularly convolutional neural networks (CNNs), are employed in precision agriculture (PA) for weed detection. This study focuses on testing CNN architectures for image classification tasks using the PyTorch framework, emphasizing hyperparameter optimization. Four groups of experiments were carried out: the first one trained all the PyTorch architectures, followed by the creation of a baseline, the evaluation of a new and extended dataset in the best models, and finally, the test phase was conducted using a web application developed for this purpose. Of 80 CNN sub-architectures tested, the MaxVit, ShuffleNet, and EfficientNet models stand out, achieving a maximum accuracy of 96.0%, 99.3%, and 99.3%, respectively, for the first test phase of PyTorch classification architectures. In addition, EfficientNet_B1 and EfficientNet_B5 stood out compared to all other models. During experiment 3, with a new dataset, both models achieved a high accuracy of 95.13% and 94.83%, respectively. Furthermore, in experiment 4, both EfficientNet_B1 and EfficientNet_B5 achieved a maximum accuracy of 96.15%, the highest one. ML models can help to automate crop problem detection, promote organic farming, optimize resource use, aid precision farming, reduce waste, boost efficiency, and contribute to a greener, sustainable agricultural future. Full article

Accurate cattle pose estimation is essential for Precision Livestock Farming (PLF). Computer vision-based, non-contact cattle pose estimation technology can be applied for behaviour recognition and lameness detection. Existing methods still face challenges in achieving fast cattle pose estimation in complex scenarios. In this work, we introduce the FasterNest Block and Depth Block to enhance the performance of cattle pose estimation based on the RTMPose model. First, the accuracy of cattle pose estimation relies on the capture of high-level image features. The FasterNest Block, with its three-branch structure, effectively utilizes high-level feature map information, significantly improving accuracy without a significant decrease in inference speed. Second, large kernel convolutions can increase the computation cost of the model. Therefore, the Depth Block adopts a method based on depthwise separable convolutions to replace large kernel convolutions. This addresses the insensitivity to semantic information while reducing the model’s parameter. Additionally, the SimAM module enhances the model’s spatial learning capabilities without introducing extra parameters. We conducted tests on various datasets, including our collected complex scene dataset (cattle dataset) and the AP-10K public dataset. The results demonstrate that our model achieves the best average accuracy with the lowest model parameters and computational requirements, achieving 82.9% on the cattle test set and 72.0% on the AP-10K test set. Furthermore, in conjunction with the object detection model RTMDet-m, our model reaches a remarkable inference speed of 39FPS on an NVIDIA GTX 2080Ti GPU using the PyTorch framework, making it the fastest among all models. This work provides adequate technical support for fast and accurate cattle pose estimation in complex farm environments. Full article

Silicon carbide wafer serves as an ideal substrate material for manufacturing semiconductor devices, holding immense potential for the future. However, its ultra-hardness and remarkable chemical inertness pose significant challenges for the surface processing of wafers, and a highly efficient and damage-free method is required to meet the processing requirements. In this study, atmospheric plasma processing was used to conduct point-residence experiments on silicon carbide wafers by varying process parameters such as Ar, CF₄, and O₂ flow rate, as well as processing power and the distance between the plasma torch and the workpiece. We investigate the effects of these on the surface processing function of atmospheric plasma etching and technique for surface modification of silicon carbide wafers, evaluating the material removal rates. Then, according to the experimentally derived influence law, suitable parameter ranges were selected, and orthogonal experiments were designed to determine the optimal processing parameters that would enable rapid and uniform removal of the wafer surface. The results indicate that the volume removal rate of the plasma on the silicon carbide wafer achieves its maximum when the input power is 550 W, the processing distance between the plasma torch and workpiece is 3.5 mm, and when the Ar, CF₄, and O₂ flow rates are 15 SLM, 70 SCCM, and 20 SCCM, respectively. Full article

A transferred arc plasma torch chemical rector was used to process waste formed from mixtures of dry clay powder and hydroquinone. Such reactors are best suited for the treatment of electrically conductive waste. In these types of reactors, the electric arc moves chaotically throughout the entire reactor volume, making it possible to ensure an even temperature distribution in the reaction zones. An analysis of the literature has shown that there are not many study results related to this type of reactor. The novelty of the work is that the behavior of the operating electric arc inside the reactor was recorded by using a high-speed camera. The distribution of the temperature profile at the cooled reactor wall was investigated. The electrical potential difference inside the reactor was also investigated. To better understand the behavioral properties of the electric arc when the reactor is filled with treated material, hydroquinone-contaminated clay was used. In this case, the movement of the electric arc, as well as the probability of its formation, is the greatest at the location where the thinnest layer of the material to be processed is located. In addition, it has been observed that the use of a graphite anode poses problems because, over time, the anode of such a design deforms due to interactions with the electric arc. While analyzing research results, it can be observed that these types of reactors are very suitable for the treatment of electrically conductive materials and for the treatment of small amounts of nonconductive materials when the material occupies a relatively small part of the reactor. A further development of these studies in the future is planned in order to make the reactors as versatile as possible and as suitable as possible for handling the widest range of materials possible. Full article

The misuse of pesticides poses a tremendous threat to human health. Excessive pesticide residues have been shown to cause many diseases. Many sensor detection methods have been developed, but most of them suffer from problems such as slow detection speed or narrow detection range. So, the development of rapid, direct and sensitive means of detecting trace amounts of pesticide residues is always necessary. A novel online sensor technique was developed for direct analysis of pesticides in complex matrices with no sample pretreatment. The portable sensor ion source consists of an MPT (microwave plasma torch) with desolventizing capability and an APCI (atmosphere pressure chemical ionization), which provides abundant precursor ions and a strong electric field. The performance which improves the ionization efficiency and suppresses the background signal was verified by using pesticide standard solution and pesticide pear juice solution measurements with an Orbitrap mass spectrometer. The limit of detection (LOD) and the limit of quantization (LOQ) of the method were measured by pear juice solutions that were obtained in the ranges of 0.034–0.79 μg/L and 0.14–1 μg/L. Quantitative curves were obtained ranging from 0.5 to 100 μg/L that showed excellent semi-quantitative ability with correlation coefficients of 0.985–0.997. The recoveries (%) of atrazine, imidacloprid, dimethoate, profenofos, chlorpyrifos, and dichlorvos were 96.6%, 112.7%, 88.1%, 85.5%, 89.2%, and 101.9% with the RSDs ranging from 5.89–14.87%, respectively. The results show that the method has excellent sensitivity and quantification capability for rapid and direct detection of trace pesticide. Full article

Deep learning–based speech recognition applications have made great strides in the past decade. Deep learning–based systems have evolved to achieve higher accuracy while using simpler end-to-end architectures, compared to their predecessor hybrid architectures. Most of these state-of-the-art systems run on backend servers with large amounts of memory and CPU/GPU resources. The major disadvantage of server-based speech recognition is the lack of privacy and security for user speech data. Additionally, because of network dependency, this server-based architecture cannot always be reliable, performant and available. Nevertheless, offline speech recognition on client devices overcomes these issues. However, resource constraints on smaller edge devices may pose challenges for achieving state-of-the-art speech recognition results. In this paper, we evaluate the performance and efficiency of transformer-based speech recognition systems on edge devices. We evaluate inference performance on two popular edge devices, Raspberry Pi and Nvidia Jetson Nano, running on CPU and GPU, respectively. We conclude that with PyTorch mobile optimization and quantization, the models can achieve real-time inference on the Raspberry Pi CPU with a small degradation to word error rate. On the Jetson Nano GPU, the inference latency is three to five times better, compared to Raspberry Pi. The word error rate on the edge is still higher, but it is not too far behind, compared to that on the server inference. Full article

The estimation of human hand pose has become the basis for many vital applications where the user depends mainly on the hand pose as a system input. Virtual reality (VR) headset, shadow dexterous hand and in-air signature verification are a few examples of applications that require to track the hand movements in real-time. The state-of-the-art 3D hand pose estimation methods are based on the Convolutional Neural Network (CNN). These methods are implemented on Graphics Processing Units (GPUs) mainly due to their extensive computational requirements. However, GPUs are not suitable for the practical application scenarios, where the low power consumption is crucial. Furthermore, the difficulty of embedding a bulky GPU into a small device prevents the portability of such applications on mobile devices. The goal of this work is to provide an energy efficient solution for an existing depth camera based hand pose estimation algorithm. First, we compress the deep neural network model by applying the dynamic quantization techniques on different layers to achieve maximum compression without compromising accuracy. Afterwards, we design a custom hardware architecture. For our device we selected the FPGA as a target platform because FPGAs provide high energy efficiency and can be integrated in portable devices. Our solution implemented on Xilinx UltraScale+ MPSoC FPGA is 4.2× faster and 577.3× more energy efficient than the original implementation of the hand pose estimation algorithm on NVIDIA GeForce GTX 1070. Full article

For better welding quality, accurate path teaching for actuators must be achieved before welding. Due to machining errors, assembly errors, deformations, etc., the actual groove position may be different from the predetermined path. Therefore, it is significant to recognize the actual groove position using machine vision methods and perform an accurate path teaching process. However, during the teaching process of a narrow butt joint, the existing machine vision methods may fail because of poor adaptability, low resolution, and lack of 3D information. This paper proposes a 3D path teaching method for narrow butt joint welding. This method obtains two kinds of visual information nearly at the same time, namely 2D pixel coordinates of the groove in uniform lighting condition and 3D point cloud data of the workpiece surface in cross-line laser lighting condition. The 3D position and pose between the welding torch and groove can be calculated after information fusion. The image resolution can reach 12.5 μm. Experiments are carried out at an actuator speed of 2300 mm/min and groove width of less than 0.1 mm. The results show that this method is suitable for groove recognition before narrow butt joint welding and can be applied in path teaching fields of 3D complex components. Full article

During the complex path workpiece welding, it is important to keep the welding torch aligned with the groove center using a visual seam detection method, so that the deviation between the torch and the groove can be corrected automatically. However, when detecting the narrow butt of a specular reflection workpiece, the existing methods may fail because of the extremely small groove width and the poor imaging quality. This paper proposes a novel detection method to solve these issues. We design a uniform surface light source to get high signal-to-noise ratio images against the specular reflection effect, and a double-line laser light source is used to obtain the workpiece surface equation relative to the torch. Two light sources are switched on alternately and the camera is synchronized to capture images when each light is on; then the position and pose between the torch and the groove can be obtained nearly at the same time. Experimental results show that our method can detect the groove effectively and efficiently during the welding process. The image resolution is 12.5 μm and the processing time is less than 10 ms per frame. This indicates our method can be applied to real-time narrow butt detection during high-speed welding process. Full article