Image fusion methods available in the literature, are designed for the office work, and are unsuitable for the use by a crew of an armored vehicle (AV). In addition, we show that the image fusion for target sightseeing system (TSS) of AVs has its own peculiarities in comparison with traditional approaches for civilian needs. We argue that for the needs of the TSS the concept of local contrast is more appropriate in comparison with the concept of integral contrast, on which the image quality indices designed for civilian needs, are based. We show that the most suitable for the needs of TSS is the fusion of multi-mode mono-temporal mono-view images synthesized from partial images at the pixel level from the visible and infrared channels. We have derived analytical expressions, which relate the local contrasts of the key points in the fused image synthesized with the simple and weighted algorithms of image fusion to the local contrast of these points in the partial images from the observation channels. From these relations we concluded that the local contrast of the image obtained via simple or weighted fusion methods is always lower than the contrast of the corresponding key points on one of the two partial images. The specific feature of the images collected from the visible and infrared channels is that contrasts of key points on such partial images, as a rule, are of opposite signs. We show that the image fusion is not always reasonable. There are situations at which the image fusion worsens the quality of the fused image in comparison with that of partial images. We demonstrate that due to the opposite signs of the local contrasts of key points on the visible and infrared images, under certain conditions the contrast of the key points in the fused image can vanish, such that the key point might become invisible in the fused image, while being clearly visible in both partial images. The analytical expressions that determine the conditions for the reduction of contrast in image fusion by simple and weight aggregation algorithms are obtained. The conclusions derived from the obtained analytical relations for the local contrast in the fused image are confirmed and illustrated by modeling the image fusion in Mathematica.