基于YOLO-v5的Ti-48Al-2Cr-2Nb合金高温析出相智能识别

Intelligent Identification of High Temperature Precipitated Phase in Ti-48Al-2Cr-2Nb Alloy Based on YOLO-v5

TiAl金属间化合物合金是一种重要的高性能材料,其微观组织对研究片层生长行为特点和组织演变规律有着重要意义。在高温相变析出生长的实验中,对其微观组织的准确识别和定位具有至关重要的作用。本文旨在解决通过人眼观察识别微观组织存在的准确率低和主观性强的问题,通过将计算机视觉技术应用于Ti-48Al-2Cr-2Nb合金高温析出相的智能识别。在经过模型选取、数据集制作、参数和配置选择以及训练结果分析后,完成了目标检测实验。结果表明,提出的析出相态组织目标检测模型在精确率、召回率和平均精度均值等指标上均能达到80%以上,其中mAP最高达到88%。同时,该方法具有高精度和高效性,单张图片的检测速度为17 ms。因此,该方法在Ti-48Al-2Cr-2Nb合金高温析出相态组织的识别和定位上具有广阔的应用前景。

TiAl intermetallic compound alloy is an important high-performance material,and its microstructure is of great significance to the study of the characteristics of lamellar growth behavior and microstructure evolution.In the experiment of high-temperature phase transition precipitation growth,the accurate identification and positioning of its microstructure plays a vital role.This paper aims to solve the problems of low accuracy and strong subjectivity in identifying the microstructure through human observation by applying computer vision technology to the intelligent identification of Ti-48Al-2Cr-2Nb alloy high-temperature precipitated phase structure.After model selection,data set production,parameter and configuration selection,and training result analysis,the target detection experiment was completed.The experimental results show that the proposed target detection model for precipitated phase structure can achieve more than 80%in terms of precision,recall and average precision,among which mAP can reach up to 88%.Meanwhile,the method has high precision and high efficiency,and the detection speed of a single image is 17ms.Therefore,this method shows broad application prospects in the identification and location of Ti-48Al-2Cr-2Nb alloy high temperature precipitated phase structure.