As the research foundation of the objective assessment,the subjective assessment of human eyes is the most direct and effective method to compare the image quality of different fusion algorithms.It is a popular research method for subjective assessment of color fused image quality at present to analyze impacts of factors as the sharpness and the color harmony on the color fused image quality based on the single dependent variable model by carrying out subjective assessment experiments,and then to predict the overall quality of fused images under vision tasks.Aiming at the problem that the method based on the single dependent variable model tends to cut apart the relationship between "detection" and "perception" in the process of detection and recognition,the vision task of detection and recognition in the village is taken as an example in this paper,"perception" and "detection" are selected as the externalization indicators of the color fused image quality.Fifteen people are organized to subjectively assess the quality of forty two color fused images generated by eight typical fusion algorithms in eight scenes.The assessment results after screening are analyzed by the structural equation model (SEM) of dual dependent variables.The prediction model of color fused image quality is established and the correctness of the model is verified.The results show that the color fused image quality under the detection and recognition vision task in the village environment can be predicted by three factors of "difference between targets and the background",and the two "scene sharpness" and "color harmony" are balanced when "perception" and "detection" are performed simultaneously.