Fig. 2 3 x 3 sharpen filter 1. The process for Implementation of fusion algorithm in spatial domain using PCA, Histogram Equalization, ny 2; ee i, ce a ee, | ee, ae oe Fig. 3 3 x 3 sharpen filter 2. Fig. 4 shows the fused results of "Single person image", where Fig. 4 (a) shows an Infrared image, Fig. 4 (b) shows a visible image, Fig. 4 (c) shows sharpen fused-image which ble images are of "Forest image" is shown in Fig. 6. The dataset used for the experiment is TNO image fusion dataset. The experiment is conducted using MATLAB software. For doing the experiment the source images uti- lized are, first set of input infrared and visible images are of "Single person image" is shown in Fig. 4. The second set of input infrared and visible images are of "Two-person image" is shown in Fig. 5. And the third set of input infrared and vis- ible images are of "Forest image" is shown in Fig. 6. The process for Implementation of fusion algorithm in spatial domain using Histogram Equalization, Sharpen infrared source image and visible source image are scaled into 256 <x 256 images. Principle Component Analysis (PCA) has been applied to the scaled infrared image and visible image. PCA is mainly used for dimensional reduc- tion. Histogram equalization is applied to the result of PCA of visible image. Histogram equalization is an enhance- ment technique that improves contrast of the image. Sharpen filter is applied to result of PCA of infrared image. Weighted average Fusion has been applied to results of his- togram equalization and sharpen filter. The 3 x 3 sharpen filter used in above method is shown in Fig. 3. average fusion rule is applied in spatial domain due to that low brightness, low contrast and less quality fused images are produced. is obtained after Implementation of DWT based fusion algorithm using sharpen filter, Fig. 4 (d) shows PCAI fused-image which is obtained after Implementation of fusion algorithm in spatial domain using PCA, Histogram Equalization, Sharpen filter and Weighted average fusion tule, Fig. 4 (e) shows PCA2 fused-image which is obtained after Implementation of fusion algorithm in spatial domain using Histogram Equalization, Sharpen filter, PCA and Weighted average fusion rule and Fig. 4 (f) shows the pro- posed fused-image of "Single person image". Similarly Fig. 5 shows the fused results of "Two-person image", where Fig. 5 (a) shows an Infrared image, Fig. 5 (b) shows a visible image, Fig. 5 (c) shows sharpen fused-im- age which is ob fusion algorithm using sharpen filter, Fig. 5 (d) shows PCA1 fused -image w ained after Implementation of DWT based hich is obtained after Implementation of fusion algorithm in spatial domain using PCA, Histogram Equa rule, after using Weig ization, S Fig. 5 (e) s mplemen posed harpen filter and Weighted average fusion hows PCA2 fused-image which is obtained ation of fusion algorithm in spatial domain Histogram Equalization, Sharpen filter, PCA and hted average fusion rule and Fig. 5 (f) shows the pro- fused-image of "Two-person image". Fused-image obtained using the proposed method is providing good contrast, texture, and edge information because of that person, house, board and tree details are clearly seen in Fig. 5 (f). Table 1 Comparison results for "Single person image" Table 2 Comparison results for "Two-person image" Table 3 Comparison results for "Forest image" Table 4 Comparison of average results of different fusion methods Fig. 8 Different fusion methods comparison with reference to Entropy Fig. 7 Different fusion methods comparison with reference to Mean Fig. 9 Different fusion methods comparison with reference to Standard deviation Fig. 10 Different fusion methods comparison with reference to Mean gradient
