激光焊接状态图像灰度共生矩阵分析法

发布时间：2018-04-18 09:02

本文选题：大功率盘形激光焊 + 金属蒸气　；参考：《焊接学报》2017年06期

【摘要】：在激光焊接过程中,金属蒸气和飞溅蕴含着丰富的焊接状态信息.以大功率盘形激光焊接304不锈钢为试验对象,应用紫外波段和可见光波段高速摄像机摄取焊接过程中金属蒸气和飞溅瞬态图像.分析图像区域纹理二阶统计特征——灰度共生矩阵,并用ASM(angular second moment)能量、惯性矩、熵和自相关性描述灰度共生矩阵,分析灰度共生矩阵与焊缝成形之间的关系,同时提取飞溅数量和面积、金属蒸气质心方位和金属蒸气高度等特征,建立BP(back propagation)神经网络模型,预测焊缝成形.结果表明,所分析方法能够有效反映金属蒸气、飞溅与焊接状态之间的关联,为在线监测大功率盘形激光焊接质量提供依据.
[Abstract]:During laser welding, metal vapor and spatter contain abundant information of welding state.High power disc laser welding 304 stainless steel was used to obtain transient images of metal vapor and splashing in the welding process by using a high speed camera in the ultraviolet and visible wave bands.This paper analyzes the second-order statistical feature of texture in image region-gray-scale co-occurrence matrix, describes the gray co-occurrence matrix with ASM(angular second moment of inertia, entropy and autocorrelation, and analyzes the relationship between gray co-occurrence matrix and weld forming.At the same time, the number and area of spatter, the orientation of metal vapor centroid and the height of metal vapor were extracted, and the BP(back propagation neural network model was established to predict the weld formation.The results show that the analysis method can effectively reflect the correlation between metal vapor, splashing and welding state, and provide the basis for on-line monitoring the quality of high-power disk-shaped laser welding.
【作者单位】：广东工业大学机电工程学院;佛山职业技术学院;
【基金】：国家自然科学基金资助项目(51675104) 广东省科技发展专项资金资助项目(2016A010102015) 广州市科技计划资助项目(201510010089) 佛山市科技创新专项资金项目(2014AG10015)
【分类号】：TG456.7

【相似文献】