Fig. 1. Sensor spectral sensitivity. (a) SS of RGB+NIR CFA; (b) SS of CMYW CFA; (c) SS of CMYW subtraction

Fig. 2. Image enhancement system architecture

Fig. 3. Experimental environment. (a) Darkroom environment for image acquisition; (b) enlargement of some details

Fig. 4. Comparison of noise reduction effects for different algorithms. (a) Difference; (b) BF; (c) JBF; (d) NLM; (e) INLM; (f) Joint NLM

Fig. 5. First group of images in different environments and images processed by different algorithms. (a) Image without NIR under normal illumination; (b) image without NIR under low illumination; (c) image with NIR under low illumination; (d) low illumination image processed by system in Ref. [11]; (e) low illumination image processed by system in Ref. [14]; (f) low illumination image processed by proposed system

Fig. 6. Second group of images in different environments and images processed by different algorithms. (a) Image without NIR under normal illumination; (b) image without NIR under low illumination; (c) image with NIR under low illumination; (d) low illumination image processed by system in Ref. [11]; (e) low illumination image processed by system in Ref. [14]; (f) low illumination image processed by proposed system

Fig. 7. Third group of images in different environments and images processed by different algorithms. (a) Image without NIR under normal illumination; (b) image without NIR under low illumination; (c) image with NIR under low illumination; (d) low illumination image processed by system in Ref. [11]; (e) low illumination image processed by system in Ref. [14]; (f) low illumination image processed by proposed system

Table 1. PSNR comparison of images processed by different noise reduction algorithms unit: dB

Table 2. Brightness mean values comparison of three groups of images

Table 3. Structural similarity comparison of three groups of images

Table 4. Comparison of PSNR of three groups of images unit: dB