天文导航技术已经进入一个新的发展阶段，随着电子技术和计算机技术的发展，它已广泛应用于卫星、航天飞机、远程导弹和其他航天器。星点质心位置是天文导航技术的关键。由于白天强烈天空背景的干扰，拍摄的近红外恒星图像信噪比很低。首先，采用信息熵的方法来进行分析恒星的图像能量分布。然后，提出了基于多级的最小能量差星点目标质心定位的方法。此方法使用线性叠加，以缩小质心区域范围，使用恒星能量分布的对称性出现的最小差值时，得到星点质心坐标。通过测试自然拍摄的图像来验证算法的准确性。实验表明：该方法具有良好的低信噪比条件下的质心定位精度。

Celestial navigation technology has entered a new stage of development with the development of electronic technology and computer technology. Otherwise it has been widely used in satellite, space shuttle, long-range missiles and other spacecraft. Star centroid location for celestial navigation is the key technologies. Due to the strong sky background, near-infrared stellar images which are taken during the day have the low signal noise ratio. First, this paper adopts the method of information entropy to make analysis the stellar images energy distribution. And then we proposed a method for star target centroid located based on multi-step minimum energy difference. This method uses the linear superposition to narrow the centroid area, using the symmetry of the stellar energy distribution to obtain the centroid when the minimum difference appears. We test the modal on natural pictures to verify the accuracy. Experiments show that it has good effect to calculate the centroid with low SNR conditions.