Objective assessment method of night vision fusion image quality

The aim of this paper is to study the practicability of the general and objective fusion image quality assessment method. Based on the analysis of the relationship between image quality assessment and fusion image quality assessment, the general expression of fusion image quality assessment was provided. Four fusion image evaluating metrics were presented based on information theory and structural similarity. Experiments were performed for 36 fusion images generated by 4 fusion algorithms fusing visible and infrared images. Experimental result shows that the objective evaluating method bonding to the human vision system is better than the metrics such as entropy and mutual information. However, the experiment result also show the assessment result is unsatisfied to the concordance subjective and objective assessing result, which explain that it is very difficult to build the fusion image evaluating metrics of the general, efficient and good consistency between subjective and objective result. Furthermore, the possible reason for performance assessment is discussed in the end.