Aiming at the features of concentrated gray-level distribution and low contrast of infrared images, a contrast enhancement algorithm based on an improved histogram equalization method is proposed. Firstly, a raw 16 bit image is mapped to 8 bit image A through linear contrast enhancement. Secondly, the raw 16 bit image is mapped to 8 bit image B by using the improved platform histogram equalization method. Thirdly, the weights of image A and image B are determined according to the gray-level range of the input images. Finally, the determined weights are used to combine image A and image B into the final contrast-enhanced image. This method has overcome the drawbacks of great noise and light mutation in traditional platform histogram equalization algorithms. Because the traditional gray-level transform enhancement algorithm is incorporated, it can adjust the contrast of the image adaptively according to the gray-level distribution of the object and background in the whole image. The experimental result shows that the proposed algorithm can keep the entire information in the image effectively while enhancing the contrast of the object.