This lecture qualification thesis investigates relevant characteristics of the human visual system. These results are used to develop methods to optimize modern visual technologies. Self-luminant displays as well as interior and exterior light sources are considered. For self-luminant displays, calibration and characterization methods are developed to display colour stimuli accurately. New display colour primaries are designed for a large colour gamut. New sub pixel architectures are proposed for an optimum spatial resolution. Modern self-luminant displays often represent large colour stimuli and their colour appearance is modelled. Chromatic afterimages are also modelled to increase the image quality of video sequences. Ergonomic principles are derived to apply colour contrast to enhance visual search performance in the presence of coloured objects. Cognitive, preferred and emotional colour is also subject of this thesis. Long-term memory colours were found to be important to enhance the perceived colour quality of colour displays. The strength of visually evoked emotions related to cinema motion pictures is computed by a mathematical model based on the technical parameters of the video sequence. The factors of colour image preference (local and global contrast, chroma and hue) are investigated. A new algorithm is shown that enhances the image quality for young and elderly observers separately. To optimize interior light sources, principles to renew the CIE colour rendering index are presented. A concept of a colour harmony rendering index is proposed. An important question of this lecture qualification thesis is the investigation of mesopic visual performance to optimize exterior light sources. To this aim, the processes of mesopic detection are modelled at the detection threshold for different retinal eccentricities. An important finding is the necessity to include the activity of chromatic retinal mechanisms in addition to the achromatic mechanisms of the current models. A new chromatic model (the CHC model) is described. The CHC model does not use the concept of luminance to model mesopic detection at or around the threshold. It models the adaptation of each mechanism of mesopic detection separately and as a function of retinal eccentricity.