超低碳钢显微组织为铁素体,在制样过程中极易出现划痕和晶界腐蚀不清晰的现象,严重影响金相组织分析。同时,显微组织特征的分析结果严重依赖于专家经验,受主观因素影响较大且效率低。为了高效获得超低碳钢显微组织特征信息,基于超低碳钢金相图像数据集,采用归一化、自适应阈值法处理图像,增强图像对比度;融合自注意力机制(Self-Attention,SA)和循环回归生成对抗神经网络(CycleGan),开发基于CycleGan+SA的晶界增强算法;建立超低碳钢显微组织特征强化模型,实现了显微组织图像的自动处理与晶界信息的特征强化。在此基础上,采用分水岭分割算法对晶界强化后的显微组织图像进行精细化分析。结果表明,CycleGan+SA算法可以有效去除原始金相图像中的划痕并补全晶界模糊区域,实现超低碳钢晶界特征强化。相比原始的CycleGan算法,引入注意力机制后,CycleGan+SA算法可以实现更清晰的晶粒分割,图像识别精确度P值由97.43%提升至98.75%,综合评价指标F值由97.49%提升至98.73%。在显微组织精细化分析方面,通过与常用分析软件对比,超低碳钢显微组织特征强化模型与Image J软件测定的晶粒尺寸平均误差为1.2个晶粒,与Image Pro Plus软件测定的晶界比例误差为0.008个百分点,模型与软件统计结果吻合较好,具备一定的应用前景。
The corrosion behavior and microstructure characteristics of metal inert gas(MIG)welded dissimilar joints of the 6005A alloy modified with Sc(designated as 6005A+Sc)and the 5083 alloy were investigated using corrosion tests and microscopy techniques.Results show that the dissimilar joints exhibit strong stress corrosion cracking(SCC)resistance,maintaining substantial strength during slow strain rate tensile tests.Notably,the heat-affected zone(HAZ)and base metal(BM)on the 6005A+Sc side show superior performance in terms of inter-granular corrosion(IGC)and exfoliation corrosion(EXCO)compared to the corresponding zones on the 5083 side.The lower corrosion resistance of the 5083-BM and the 5083-HAZ can be attributed to the presence of numerous Al_(2)Mg_(3)phases and micro-scaled Al_(6)(Mn,Fe)intermetallics,mainly distributed along the rolling direction.Conversely,the enhanced corrosion resistance of the 6005A+Sc-BM and the 6005A+Sc-HAZ can be attributed to the discontinuously distributed grain boundary precipitates(β-Mg_(2)Si),the smaller grain size,and the reduced corrosive current density.
