Characteristic fusion-based electric power inspection image rapid alignment method

Based on an SIFT algorithm and an RANSAC algorithm, the method specifically comprises the following steps: based on the SIFT algorithm, improving a feature descriptor of the SIFT algorithm, generating a feature descriptor by using an LATCH method, forming a bit string by the LATCH descriptor through calculating a pixel block comparison value in a window, and carrying out feature extraction on the bit string; determining the value of each bit of the bit string by comparing the gray values of the pixel point pairs in the related region of the feature points; based on an RANSAC algorithm, a self-adaptive minimum threshold method and self-adaptive minimum sampling times are adopted for calculation. According to the method, the operation speed in the rapid image alignment process is increased, the problem of low operation efficiency is solved, and meanwhile the problem of the upper limit of the number of iterations of the algorithm and the problem of result errors are solved. 本发明公开了一种特征融合的电力巡视图像快速对准方法，基于SIFT算法和RAN