Creating a living environment that is both comfortable and human-centric is of paramount importance for the sustainable development of cities. Traditional methods of studying urban perceptions, which typically involve field surveys, questionnaire surveys, and remote sensing inversion, often fall short when it comes to accurately measuring and portraying the perceptual experiences of residents in relation to their visual environment. To address this shortcoming, a new framework has been developed, grounded in the theory of Landsenses ecology. This framework provides a quantitative measurement and coupling analysis of urban visual perception, encompassing both physical and psychological aspects. The primary aim of this framework is to gauge the subjective feelings of individuals towards the built environment from a human perspective. This innovative framework offers several key advantages. Firstly, it supplements existing methods of analyzing urban micro-space patterns, providing a more comprehensive understanding of the urban environment. Secondly, it enriches our understanding of changes in physical and psychological perceptual space from a human perspective. This is achieved by enhancing our ability to identify spatial patterns and detail depiction of urban perceptual quality. Lastly, it deepens our understanding of the complex relationships between different elements of urban composite ecosystems and residents' perceptions. The center of Beijing serves as a case study for the application of this framework. A set of deep learning approaches has been developed to quantify physical and psychological perception measurements at the human-scale. Following this, a spatial statistical model is employed to identify the intensity of the effects of physical perception and urban composite ecosystem elements on residents' psychological perception. The findings of this study are illuminating. Firstly, it was found that the visibility of trees (TREE) and the normalized difference vegetation index (NDVI) play a significant role in enhancing positive psychological perception. Secondly, a notable spatial clustering phenomenon was observed in the psychological perception index (PPI) in the center of Beijing. This phenomenon exhibited significant differences in the "inner-outer" layers and "south-north" urban areas. Lastly, it was demonstrated that deep and machine learning models can be effectively used to accurately depict the spatial patterns of urban physical and psychological perceptions. In conclusion, this study provides a valuable quantitative reference for the creation of landsenses creation and ecological planning at the urban block scale. It holds significant implications for the intelligent management of sustainable urban development, offering a novel perspective on how we understand and shape our urban environments and human well-beings.