Abstract: White cap is a very common phenomenon at ocean surface, and accurate identification and extraction of it is ctritically important to the study of air-sea interaction and marine remote sensing. Automatic white cap extraction technology based on digital image processing has the characteristics of high processing efficiency, good accuracy, and low cost. However, the images taken under real sea condition are usually affected by solar illumination, which may lead to false white cap extraction and make the accurate extracting of white cap coverage (W_C) much difficult. This may also has significant impact on the parameterization of white cap coverage. In this paper, based on the traditional adaptive threshold segmentation algorithm, an improved light correction algorithm considering homomorphic filtering and image enhancement is proposed for wave images with uneven illumination to eliminate the influence of uneven illumination on white cap recognition. The algorithm is then verified with the ocean wave images captured by a stereo photogrammetry system. Compared with the results of traditional white cap extraction algorithm, both the accuracy of white cap extraction under uneven illumination and the calculation accuracy of white cap coverage are greatly improved, suggesting that the new algorithm can effectively improve the accuracy and reliability of the fitting relation between white cap coverage and 10 meters sea surface wind speed.