提出了一种基于场景特征点的双目变焦系统相机主距的计算方法。首先选取任意焦距作为基础焦距，用常规方法获取基础焦距下双目系统的内外参数，利用ASIFT算法提取基础焦距下场景中待测物体的特征控制点，利用视差原理重建空间世界点，接着切换焦距后再次提取两个焦距下待测物体对应的左右图像特征点，将不同焦距下匹配的图像特征点作为共同控制点，最后根据CCD平面与待测物体空间位置不变性完成双目变焦相机的主距求解。实验结果表明，该算法实际操作简单，主距计算结果的平均相对误差为2.1%。

A method for calculating the principal distance of a binocular zoom system is presented based on scene feature points. First of all, an arbitrary focal length is selected as the basis of focal length, which is utilized to obtain internal and external parameters of the binocular system by traditional method. Then the feature points of the basis-focal-length-based scene can be extracted through ASIFT algorithm before reconstructing the 3D points. After that, the focus and extracting the feature points of left and right images of two focus situations are changed before matching feature points, which are regarded as the common control points. At last, the principle distance of binocular zoom lens cameras are obtained based on the invariant space position of CCD and objects. The experimental results show that the algorithm is simple, and the average relative error of the result is 2.1%.