Lodging resistance of wheat is closely related to its stem section microstructure. Observation and analysis of stem microstructure are of great significance for the genetic breeding. A microscopic image measurement system was studied, which could be utilized to measure the main parameters of the wheat stem section accurately. The equipment including computer platform, image acquisition box and electronic microscopy imaging system etc. was used to build the hardware platform and the algorithm software was developed. According to the characteristics of the stem structure and color, the contours of both mechanical tissue and medullary cavity were obtained using image processing procedures. After then the geometrical parameters of the whole section, medullary cavity and sclerenchyma were calculated by the ellipses fitting. Based on the uniformity of the vascular bundle structure, an original template matching method was proposed. This method was invariant to scaling, translation and rotation, which was used to identify and count the vascular bundle in the target image. It showed that this automated measurement system was of high throughput and high precision. The system will play an important role in both the screening of important crops resources and population genetic research.