Aiming at the problem of highefficiency visual attention mechanism in image target recognition, a culture image target detection algorithm based on visual saliency model and image segmentation is proposed. Firstly, the existing visual significant calculation method is improved based on the color component contrast visual saliency model, and the efficiency of significant area detection could be improved. Then, in the principle of the traditional GrabCut image segmentation algorithm, combined with the improved visual saliency model to achieve significant area initialization, no manual interaction steps are required, thus achieving fully automatic image target detection. The test data shows that compared with the existing typical methods, the proposed algorithm has better target detection performance and faster detection speed.