The accurate human detection based on artificial intelligence plays a crucial role in various applications such as selfdriving and augmented reality. In poor lighting conditions, infrared images are more suitable than visible light images. However, infrared images-based human detection has many shortcomings and improving infrared images-based human detection is still a fundamental task. In this study, three types of pre-processing are applied to infrared images to improve the accuracy of deep learning-based human detection. To further improve the accuracy, we propose an image fusion algorithm using both visible and infrared images. Experimental results show that the proposed method performs better than conventional methods.