Design of a Garden Plant Community Landscape Evaluation Method Based on the SBE Approach

In urban ecology and landscape architecture, the objective evaluation of garden plant community landscapes remains a significant challenge, primarily due to the subjective nature of traditional assessment methods and their limited capacity to handle large-scale data. This study introduces a novel approach based on the Scene-Based Evaluation (SBE) method, leveraging Convolutional Neural Networks (CNNs) to extract key features such as vegetation coverage and species diversity indexes from landscape images. Grounded in principles of ecology and landscape aesthetics, we have developed a comprehensive set of evaluation indicators to assess the ecological and aesthetic values of garden plant communities systematically. Our method significantly outperforms existing evaluation techniques, as demonstrated through extensive testing across multiple datasets. This paper details the theoretical framework, modeling steps, and the effectiveness of the proposed approach, offering a new direction for objective and scalable landscape evaluation in the context of garden plant communities. By integrating advanced image processing techniques with ecological and aesthetic principles, this study contributes a valuable tool for researchers and practitioners in landscape assessment, paving the way for more sustainable urban and garden landscape planning and management.