To obtain and make full use of the multi-spectrum information of the scene， and to lay the foundation for the research of multi-band image algorithms， a four-band coaxial imaging experimental platform of visible light + near infrared， short wave infrared， mid-wave infrared and long-wave infrared was developed based on uncooled infrared focal plane array （IRFPA）. Image processing cases based on the four-band imaging experimental platform were presented. FPGA was used in the preprocessing of each channel of image and the fusion of multi-band image. The image quality of each band was optimized by the preprocessing procedure， that included the blind point correction of the visible sensor and nonuniformity correction of the mid-wave and long-wave infrared sensors based on grad-filtering. Dual-band and four-band image fusion enhanced the color of the fusion image obtained by linear combination in the YUV color space by color transfer， thus resulting in a natural color fusion image as well as ensuring the operating efficiency of the algorithm. Dual-band infrared temperature measurement was performed using medium and long-wave infrared. The blackbody experiment shows that in the temperature range of 20-80 ℃， the accuracy of dual-band temperature measurement is significantly higher than that of single-band temperature measurement， the error of dual-band temperature measurement is less than 3.5%， and the single-point temperature fluctuation range is less than 0.7%. The outfield experiment results show that information fusion images can effectively enhance the human eye’s understanding of scene information. The four-band imaging experimental platform can easily obtain the four-band information of the scene， and image fusion and dual-band temperature measurement can strengthen the understanding of scene information.