首先采用传统定量金相方法完成两种新型铁路车轮钢中先共析铁素体占比的计算，然后采用MATLAB软件对车轮钢组织的SEM图进行灰度处理，观察到其灰度直方图为单峰分布。结合车轮钢组织SEM图的灰度直方图特点，提出了一种适合于车轮钢组织SEM图分割的改进三角形算法，算法主要思想为：连接灰度直方图波形中波峰最大值和左侧最小值形成线段，作垂直于该线段并与波形相交距离最长的垂线，垂线与直方图波形交点所对应的阈值即为选取的图像分割阈值。与传统定量金相分析方法结果相比，该方法具有较好的精确度，计算效率高，可以为计算铁路车轮钢中先共析铁素体相-珠光体相的组织比例提供借鉴。

Firstly, the traditional quantitative metallographic method was used to calculate the proportion of proeutectoid ferrite in two new railway wheel steel. Secondly, the grey treatment for the SEM image of the wheel steel structure was carried out by using MATLAB software. It was observed that the grayscale histogram was a single peak distribution. Afterwards an improved triangle algorithm for segmentation of SEM image of wheel steel structure was proposed based on the characteristics of gray histogram. The main idea of the algorithm is as follows: a line segment could be acquired by connecting the maximum value of the wave crest and the minimum value on the left side of the gray histogram waveform, then a vertical line to the acquired line segment would be made and the intersection distance to the waveform should be the longest. Furthermore, the ultimate intersection point between the vertical line and the histogram waveform would be selected as the image segmentation threshold. Compared with the results of traditional quantitative metallographic analysis method, this method had better accuracy and higher calculation efficiency, which can provide reference for calculating the proportion of proeutectoid ferrite phase and pearlite phase in railway wheel steel.