Real-time detection of marine infrared objects based on the fusion of gradient information

A feasible method considering the character of marine infrared image was proposed to detect objects in the sequential images from surface vehicle, which was not only appropriate for sea-sky background but also for offshore background. There was no need to filter the noise. Firstly, the complexity of sub-images and the average gray difference of their up and down neighborhood were measured to predict the sea line region. Secondly, improved Canny edge detection was applied to extract the contour of the sea line region. It made the sea line obvious and meaningless edges disappear greatly. Thirdly, Hough transformation was used to pick the longest line as the sea line. Finally, a kind of general concept of gradient was put forward. The targets could be marked excellently under the fusion of gradient. The experiment results show that this method can locate the sea-line region and the sea line with any tilt reliably and obtain the information of objects effectively. The whole procedure costs about 60 ms and it is real-time and robust.