Improved image fusion algorithm for detecting obstacles in forests

The harvester has been widely used in forestry operations. However the existence of obstacles in forests, such as stones, animals, human and clustered trees, may leads to safety accident and decreases the operating efficiency of the harvester. Therefore it becomes vital to show the obstacles accurately to provide guidance for the harvester. In this paper, an improved image fusion algorithm based on Contourlet transform and pulse coupled neural network (PCNN) is proposed to enhance the fused images' clarity and capture more abundant information of the reality. As for comparison, algorithms, such as wavelet transform, Principal Component Analysis (PCA) and PCNN are simulated to evaluate the proposed algorithm. Visible and infrared thermal image are captured in forest in this experiment. The experimental results demonstrate that the proposed algorithm outperform the mentioned methods by the objective criteria of Entropy, Average gradient and Standard deviation.