基于智能机器人视觉清理系统研究

针对铸造类零件人工清理效率低、劳动强度大、成本高等问题,论文提出一种基于机器视觉的智能机器人清理系统。设计了由CCD摄像机、图像采集卡、输入输出单元、控制装置构成的机器人硬件系统,搭建了图像采集、处理分析、学习识别软件系统。通过对金属型铸造的盖板、导体类零件浇口、冒口等清理部位机械化自动化清理工作的多次实验验证,零件残留小于1mm,清理准确率达到97.6%,该系统所有指标达到设计要求,满足了工厂对于零部件毛刺智能化清理的要求。     

下沟煤矿本煤层预抽钻孔合理封孔长度研究

封孔长度的合理性对于煤层预抽钻孔的抽采质量影响较大，当封孔长度较大时，不利于封孔段瓦斯的抽采，同时封孔成本增加，反之当封孔段长度较小时，若未封孔部分与煤壁裂隙导通，钻孔发生漏气又影响钻孔抽采质量。为确定合理的封孔长度，针对下沟煤矿302工作面特定的地质条件，采用单孔声波发射对巷道卸压区和应力集中区范围进行了研究，通过分析卸压区范围，确定了合理的封孔长度，并结合试验孔周边的预抽钻孔抽采数据，进行了验证分析，得出该工作面的合理封孔深度为8 m。     

卡头零件多工位级进模设计

根据卡头零件的结构特点及实际生产的技术要求,进行多工位级进模排样设计,由于卡头零件结构为对称结构,故其排样设计采用对称单排结构设计,导正孔设计采用零件本身处于对称轴处的工艺孔,以保证板料冲裁、折弯时的强度,以及确保送料精度,同时采用导正销确定步距。该卡头零件多工位级进模采用顺装结构,使得工件与废料自然分离,有利于减少人工成本。同时,由于该零件折弯次数较多,为保证板料成形精度,故合理地进行了空工位的设计,避免成形零件发生回弹等影响零件精度。生产结果表明,卡头零件多工位级进模能满足产品生产要求,该模具的材料利用率高达75.62%,极大地节省了材料。     

A Comparative Study of Physicochemical and Flavor Characteristics of Chicken Nuggets during Air Frying and Deep Frying

Chicken nuggets were air-fried and deep-fried for various times at the same temperature. The moisture content, oil content, color, texture, volatile compounds, and lipids degradation of the chicken nuggets were investigated. The evolution rate of the moisture content, oil content, color, and texture were much higher during deep frying than during air frying. Air frying required more time to produce similar quality attributes as those of deep-fried chicken nuggets. On average, the oil content of the air-fried chicken nuggets was 25% lower than that of the deep-fried chicken nuggets. Of the 74 volatile compounds identified in the fried chicken nuggets, aldehydes, hydrocarbons, and heterocycles were the main flavor contributors. The total amount of volatile compounds and aldehydes detected in the air-fried chicken nuggets was much lower than that in the deep-fried samples, but the number of heterocycles produced by the Maillard reaction was higher during air frying. Sensory evaluation revealed that the air-fried chicken nuggets had lower oiliness and crispness, and 9 min was the optimal frying time to obtain air-fried chicken nuggets with similar sensory attributes (color, odor, juiciness, and overall acceptance) as those of the best deep-fried chicken nuggets. Air frying was a healthier frying method for the production of fried chicken nuggets, which could reduce oxidative degradation of lipids in chicken nuggets.     

A tutorial on the segmentation of metallographic images: Taxonomy,new MetalDAM dataset,deep learning-based ensemble model,experimental analysis and challenges

Image segmentation is an important issue in many industrial processes, with high potential to enhance the manufacturing process derived from raw material imaging. For example, metal phases contained in microstructures yield information on the physical properties of the steel. Existing prior literature has been devoted to develop specific computer vision techniques able to tackle a single problem involving a particular type of metallographic image. However, the field lacks a comprehensive tutorial on the different types of techniques, methodologies, their generalizations and the algorithms that can be applied in each scenario. This paper aims to fill this gap. First, the typologies of computer vision techniques to perform the segmentation of metallographic images are reviewed and categorized in a taxonomy. Second, the potential utilization of pixel similarity is discussed by introducing novel deep learning-based ensemble techniques that exploit this information. Third, a thorough comparison of the reviewed techniques is carried out in two openly available real-world datasets, one of them being a newly published dataset directly provided by ArcelorMittal, which opens up the discussion on the strengths and weaknesses of each technique and the appropriate application framework for each one. Finally, the open challenges in the topic are discussed, aiming to provide guidance in future research to cover the existing gaps.     

双面板非金属孔内渗油研究

双面板非金属孔内渗油是PCB感光阻焊制程中最常见的品质缺陷之一。就双面板非金属孔内渗油的品质问题进行了研究,设计试验对各影响因素进行分析,确定双面板非金属孔内渗油的根本原因及改善方向。     

Emotion recognition at the edge with AI specific low power architectures

Nowadays Deep Learning is applied in almost every research field and helps getting amazing results in a great number of challenging tasks. The main problem is that this kind of learning and consequently Neural Networks that can be defined deep, are resource intensive. They need specialized hardware to perform computation in a reasonable time. Many tasks are mandatory to be as much real-time as possible . It is needed to optimize many components such as code, algorithms, numeric accuracy and hardware, to make them “efficient and usable”. All these optimizations can help us to produce incredibly accurate and fast learning models. The paper reports a study in this direction for the challenging face detection and emotion recognition tasks.     

Pathways to controlled 3D deformation of graphene: Manipulating the motion of topological defects

Functional properties of 2D materials like graphene can be tailored by designing their 3D structure at the Angstrom to nanometer scale. While there are routes to tailoring 3D structure at larger scales, achieving controllable sub-micron 3D deformations has remained an elusive goal since the original discovery of graphene. In this contribution, we summarize the state-of-the-art in controllable 3D structures, and present our perspective on pathways to realizing atomic-scale control. We propose an approach based on strategic application of mechanical load to precisely relocate and position topological defects that give rise to curvature and corrugation to achieve a desired 3D structure. Realizing this approach requires establishing the detailed nature of defect migration and pathways in response to applied load. From a computational perspective, the key needed advances lie in the identification of defect migration mechanisms. These needed advances define new forward and inverse problems: when a fixed stress or strain field is applied, along which pathways will defects migrate?, and vice versa. We provide a formal statement of these forward and inverse problems, and review recent methods that may enable solving them. The forward problem is addressed by determining the potential energy surface of allowable topological configurations through Monte Carlo and Gaussian process models to determine defect migration paths through dynamic programming algorithms or Monte Carlo tree search. Two inverse models are suggested, one based on genetic algorithms and another on convolutional neural networks, to predict the applied loads that induce migration and position defects to achieve desired curvature and corrugation. The realization of controllable 3D structures enables a vast design space at multiple scales to enable new functionality in flexible electronics, soft robotics, biomimetics, optics, and other application areas.     

Deep reinforcement learning-based safe interaction for industrial human-robot collaboration using intrinsic reward function

Aiming at human-robot collaboration in manufacturing, the operator's safety is the primary issue during the manufacturing operations. This paper presents a deep reinforcement learning approach to realize the real-time collision-free motion planning of an industrial robot for human-robot collaboration. Firstly, the safe human-robot collaboration manufacturing problem is formulated into a Markov decision process, and the mathematical expression of the reward function design problem is given. The goal is that the robot can autonomously learn a policy to reduce the accumulated risk and assure the task completion time during human-robot collaboration. To transform our optimization object into a reward function to guide the robot to learn the expected behaviour, a reward function optimizing approach based on the deterministic policy gradient is proposed to learn a parameterized intrinsic reward function. The reward function for the agent to learn the policy is the sum of the intrinsic reward function and the extrinsic reward function. Then, a deep reinforcement learning algorithm intrinsic reward-deep deterministic policy gradient (IRDDPG), which is the combination of the DDPG algorithm and the reward function optimizing approach, is proposed to learn the expected collision avoidance policy. Finally, the proposed algorithm is tested in a simulation environment, and the results show that the industrial robot can learn the expected policy to achieve the safety assurance for industrial human-robot collaboration without missing the original target. Moreover, the reward function optimizing approach can help make up for the designed reward function and improve policy performance.     

Evaluation of the effect of a novel membrane filtration system on the lifespan of frying oil

Frying oil has a limited lifetime which results in the disposal of approximately 67% of all used oil. The aim of this study is to determine the suitability of different membranes for purifying frying oil and investigate the application of continuous membrane filtration in a commercial deep fryer. Eleven different membranes were tested to assess their suitability for filtrating frying oil. A prototype continuous filtration system with an integrated membrane filter was developed for a deep fryer to assess the effect of continuous membrane filtration on oil degradation. The prototype was subjected to a frying test over 13 days and the results compared to those from a standard deep fryer. The prototype maintained total polar compounds (TPC) at an acceptable level of <12% and the oil remained light in colour even after more than 280 h of use, while TPC in the standard deep fryer rose to >27%.