A color correction algorithm for four-band (the RGB bands and an additional near-infrared spectrum) tree image based on RGB color space is proposed to correct the images which influenced by near-infrared spectrum. Firstly, based on the difference of R band and G band in the image, obtaining the best binarization threshold of four-band image through the experiment. Second, using the median filter and mathematical morphology to de-noise the binary images, and extracting the area which severely polluted by the near-infrared spectrum. At last, correcting images in different areas with different method, eliminating near-infrared pollutions in the image. The experimental results show that, the algorithm has good correction effect for the images which include plants.