Abstract: To address the issues of strong subjectivity, high labor intensity and limited detection methods of morphological identification and parameter analysis of tobacco stems in cigarette rods by traditional manual detection, a method based on computed tomography (CT) detection and imaging technology was proposed and evaluated for this purpose. By employing the filtered back projection method for three-dimensional reconstruction and conducting gray threshold segmentation and connected component data treatments, key stem parameters such as the length, volume and quantity inside cigarette rods were deduced accurately and efficiently. The results showed that: 1) CT imaging for Q brand cigarette detection achieved a RSD less than 1.00% with an mean absolute error of 0.15 mm for stem length, while a RSD less than 5.00% for spatial position detection, demonstrating high accuracy and excellent repeatability. 2) For P, L and Y brand cigarettes, the accuracy rates for both stem quantity and morphology identification reached near 100%, demonstrating excellent repeatability and consistency. This study provides a novel non-destructive testing technique for morphology identification and parameter analysis of tobacco stems in cigarettes.