面料图像属性描述及数据库设计

由于现有对面料或面料图像属性描述的研究或应用各不相同,没有一个完整的标准可以遵循。为了研究面料图像的属性描述及数据库设计,基于情感感知的面料图像数据库系统设计与开发,通过对现有文献的研究和分析,根据基于情感感知数据库系统的实际需要,结合相关的研究成果,对面料图像的属性描述进行研究,并根据这些描述对数据库进行设计,为基于情感感知的数据库系统软件开发提供基础。

结合面料属性和触觉感测的织物识别

目的织物识别是提高纺织业竞争力的重要计算机辅助技术。与通用图像相比,织物图像通常只在纹理和形状特征方面呈现细微差异。目前常见的织物识别算法仅考虑图像特征,未结合织物面料的视觉和触觉特征,不能反映出织物本身面料属性,导致识别准确率较低。本文以常见服用织物为例,针对目前常见织物面料识别准确率不高的问题,提出一种结合面料属性和触觉感测的织物图像识别算法。方法针对输入的织物样本,建立织物图像的几何测量方法,量化分析影响织物面料属性的3个关键因素,即恢复性、拉伸性和弯曲性,并进行面料属性的参数化建模,得到面料属性的几何度量。通过传感器设置对织物进行触感测量,采用卷积神经网络(convolutional neural network,CNN)提取测量后的织物触感图像的底层特征。将面料属性几何度量与提取的底层特征进行匹配,通过CNN训练得到织物面料识别模型,学习织物面料属性的不同参数,实现织物面料的识别并输出识别结果。结果在构建的常见服用织物样本上验证了本文方法,与同任务的方法比较,本文方法识别率更高,平均识别率达到89.5%。结论提出了一种基于面料属性和触觉感测的织物图像识别方法,能准确识别常用的服装织物面料,有效提高了织物识别的准确率,能较好地满足实际应用需求。

探析服装面料的属性在服装款式设计中的应用与创新

不同的面料与服装款式搭配,都会产生不同的视觉效果.对面料属性的研究就变得尤为重要,也成为服装设计师需要掌握的必备设计技能.

Materials flow and phase transformation in friction stir welding of Al 6013/Mg

Material flow and phase transformation were studied at the interface of dissimilar joint between Al 6013 and Mg, produced by stir friction welding （FSW） experiments. Defect-free weld was obtained when aluminum and magnesium were placed in the advancing side and retreating side respectively and the tool was placed 1 mm off the weld centerline into the aluminum side. In order to understand the material flow during FSW, steel shots were implanted as indexes into the welding path. After welding, using X-ray images, secondary positions of the steel shots were evaluated. It was revealed that steel shots implanted in advancing side were penetrated from the advancing side into the retreating side, whereas the shots implanted in the retreating side remained in the retreating side, without penetrating into the advancing side. The welded specimens were also heat treated. The effects of heat treatment on the mechanical properties of the welds and the formation of new intermetallic layers were investigated. Two intermetallic compounds, Al3Mg2 and Al12Mg17, were formed sequentially at Al6013/Mg interface.

Fabrication of cast carbon steel with ultrafine TiC particles

The carbon steels dispersed with ultrafine TiC particles were fabricated by conventional casting method. The casting process is more economical than other available routes for metal matrix composite production, and the large sized components to be fabricated in short processing time. However, it is extremely difficult to obtain uniform dispersion of ultrafine ceramic particles in liquid metals due to the poor wettability and the specific gravity difference between the ceramic particle and metal matrix. In order to solve these problems, the mechanical milling (MM) and surface-active processes were introduced. As a result, Cu coated ultrafine TiC powders made by MM process using high energy ball milling machine were mixed with Sn powders as a surfactant to get better wettability by lowering the surface tension of carbon steel melt. The microstructural investigations by OM show that ultrafine TiC particles are distributed uniformly in carbon steel matrix. The grain sizes of the cast matrix with ultrafine TiC particles are much smaller than those without ultrafine TiC particles. This is probably due to the fact that TiC particles act as nucleation sites during solidification. The wear resistance of cast carbon steel composites added with MMed TiC/Cu-Sn powders is improved due to grain size refinement.

Two Kinds of Calculation Methods in Whole Process and a Best New Comprehensive Figure of Metallic Material Behaviors

Based on the viewpoint that there are similarities to the genetic principles in the mechanics and the engineering fields; based on data in some referents and author＇s theoretical approach, this paper is to consider the microscopic-damage and the macroscopic-damage behaviors of materials that are distinct differences, thereby to propose computing models and methods of subsection calculations in whole process; to consider the behaviors between the short cracks and long ones both of which are always continuous, thereby to propose some computing models and methods of the successive calculations. These computing models refer to formulas of the threshold sizes of cracks （or threshold values of damages）; the propagating rates of cracks; the predicting calculations of lifetime; Particularly, in which it provides a best new comprehensive figure that it could be including mid describing all problems mentioned above. So it may be as a bridge to link the traditional material mechanics, the material discipline and the modem mechanics on fatigue-damage-fracture; perhaps, it can also be as route diagram to guide designs and calculations to some materials and structures. Therefore, above works realize calculations of the strength problems, the growth rate of cracks （damages） mid prediction of lifetime in whole process that would have practical significances.

Surface effects of liquid metal amoeba

Liquid metals(LM) such as eutectic gallium-indium and gallium-indium-tin are important functional liquid-state metal materials with many unique properties, which have attracted wide attentions especially from soft robot area. Recently the amoeba-like transformations of LM on the graphite surface are discovered, which present a promising future for the design and assemble of self-fueled actuators with dendritically deformable body. It appears that the surface tension of the LM can be significantly reduced when it contacts graphite surface in alkaline solution. Clearly, the specific surface should play a vital role in inducing these intriguing behaviors, which is valuable and inspiring in soft robot design. However, the information regarding varied materials functions underlying these behaviors remains unknown. To explore the generalized effects of surface materials in those intriguing behavior, several materials including glass, graphite, nickel and copper oxides(CuO) were comparatively investigated as substrate surfaces.Important results were obtained that only LM amoeba transformations were observed on graphite and CuO surfaces. In order to identify the proper surface condition for LM transformation, the intrinsic properties of substrate surfaces, such as the surface charge and roughness, as well as the specific interaction with LM like wetting behavior and mutual locomotion etc., were characterized. The integrated results revealed that LM droplet appears more likely to deform on surfaces with higher positive surface charge density, higher roughness and less bubble generation on them. In addition, another surface material,CuOx, is identified to own similar ability to graphite, which is valuable in achieving amoeba-like transformation. Moreover, this study offers a fundamental understanding of the surface properties in realizing LM amoeba transformations, which would shed light on packing and structure design of liquid metal-based soft device within multi-material system.