In this study, a new quantifiable and refined urban street color analysis method was proposed by combining professional color cards and efficient software color recognition, which solved the problems of low efficiency and difficulty in the quantification of urban color research and analysis. The research mainly uses China Building Color Card (CBCC) and Python (use programs for the HSV color segmentation of pictures) and other software to carry out color recognition for a street view. From the aspects of color composition, type, proportion, visual level, and color sequence of the street facade, this article makes a quantitative analysis of the color of Avenida de Almeida Ribeiro in Macao from multiple angles. The method of combining color card colorimetry with computer color recognition, which not only considers the inherent color of the building but also includes the color situation under the influence of the environment, can express the “actual color situation” of the building more completely. This article quantifies, combs, summarizes, and compares architectural color and environmental color completely. This method has good universality and ease of use in practice, and the conclusion of the study can provide a reference for the color planning of Macao, the color selection of urban renewal has reference significance, and provide a new method for the study of urban color.

The objective of this inquiry is to illuminate the attributes of heavy metal contamination and evaluate the potential ecological hazards inherent in the surface soil of Chenzhou City. A comprehensive analysis was conducted on 600 systematically collected soil samples within the study area, utilizing enrichment factors, geo-accumulation indices, comprehensive pollution indices, potential ecological hazard indices, and health risk assessment models to evaluate the degree of heavy metal contamination in the soil, potential ecological risks, and associated health hazards. The findings reveal that the average enrichment factor (EF) for each heavy metal is below 2, with the hierarchy from highest to lowest being Hg > Cd > Cu > Pb > Ni > Zn > Cr > As. Approximately 78.67% of soil samples exhibit no pollution to weak pollution levels based on heavy metal enrichment factors. Moreover, the comprehensive pollution index (IPI_N) indicates that 95.17% of soil samples are within safe and pollution-free levels, indicating an overall environmentally secure setting. However, 2.67% of samples display heightened potential ecological risk levels, primarily concentrated in the southwestern region of the study area, influenced by nearby industrial activities. Additionally, it is noteworthy that both the non-carcinogenic and carcinogenic health hazards emanating from soil heavy metals to adult individuals lie within tolerable thresholds. Among these, arsenic (As), chromium (Cr), and lead (Pb) have been discerned as the principal non-carcinogenic agents. It is of particular significance that only a solitary soil specimen, located in the southwestern quadrant of the investigative region, manifests detectable health perils for children.

This paper takes the garden road system as the research object and proposes a method of generating paths for classical gardens based on parametric design. Firstly, by studying the distribution characteristics of roads, the data on the curvature, angle, and view area of roads were collected. Secondly, the obtained data were transferred to the parameterized platform, and a method of intelligent generation was used for calculation. Finally, the road system was optimized by the genetic algorithm for better application in modern landscape design. According to the current situation, the road system plan generated by the algorithm inherits the characteristics of classical garden roads. This method can be applied to the courtyard, the community park, the urban park, and other objects. This research not only identifies the characteristics of landscape cultural heritage, but also produces an innovative, intelligent design tool. It provides new methods for the parameterized inheritance and application of traditional landscape heritage.

We propose a hierarchical federated learning (HFL) framework for predictive pollutant analysis in advanced green analytical chemistry (AGAC), addressing the limitations of centralized approaches in scalability and data privacy. The system integrates localized sub-models with hybrid neural architectures, combining LSTM and attention mechanisms to capture temporal dependencies and feature importance in distributed analytical data, while raw measurements remain decentralized. A global aggregator dynamically adjusts model weights based on validation performance and data heterogeneity, ensuring robust adaptation to diverse environmental conditions. The framework interfaces seamlessly with AGAC infrastructure, processing inputs from analytical instruments into standardized sequences and mapping predictions back to pollutant concentrations through calibration curves. Implemented with PyTorch Federated and edge-cloud deployment, the system employs homomorphic encryption for secure data transmission, prioritizing spectral features critical for organic pollutant detection. Our approach achieves superior accuracy and privacy preservation compared to traditional centralized methods, offering a transformative solution for scalable environmental monitoring. The proposed method demonstrates significant potential for real-world applications, particularly in scenarios requiring distributed data collaboration without compromising analytical integrity.

Lingnan gardens represent one of the three major schools of classical Chinese gardens. As the manifestation of regional culture and ecological wisdom in South China, they are considered irreplaceable in terms of their inherited historical context and role in shaping the spirit of locations. The aim of this study was to investigate the public’s perception of the commonalities and characteristics of Lingnan gardens as well as group differences in visual perception to accurately understand the public’s visual preferences for Lingnan gardens and promote their protection, utilization, and innovative development. A convenience sampling method was used to the semantic differential (SD) method, and Yuyin Garden in Guangzhou, Foshan Liangyuan Garden, Dongguan Keyuan Garden, and Shunde Qinghui Garden were selected as sample collection locations. A total of 120 valid data points were collected. Factor analysis, descriptive statistics and chi-square test were used to explore the perceptual dimensions and inter-group differences. The findings of the study were as follows: (1) Principal component factor analysis revealed that the four factors with the greatest impact on public visual perception were visual attention factors, visual information content factors, field of vision factors, and spatial visual factors. (2) Among the evaluation factors in the 120 questionnaires, the overall scoring trend was consistent, and the visual perception results of Lingnan gardens in different locations were similar. Natural elements received the highest evaluation score, and the public’s evaluation of this issue was the most unified. The evaluation of area size and Western cultural elements had the most significant differences, and the opinions of the respondents were difficult to unify. (3) There were two items that differed due to different groups of people and were divided into two categories: growth factors and non-growth factors, among which growth factors have a greater impact. The differences in visual perception of Lingnan gardens were explored in this study from the perspective of the public, and a quantitative translation path for analyzing the differences in visual perception of classical Chinese gardens is proposed. Although the sample size is sufficient for an initial understanding, it is recommended that future studies involve larger and more stratified samples to enhance generalizability.

We propose a novel hybrid architecture that integrates a continuous-time Liquid Neural Network with an optimized Long Short-Term Memory network for seismic vulnerability state assessment of ancient halls. The proposed system processes multi-source structural monitoring data, including acceleration, strain, displacement, temperature, and humidity signals. A central challenge in this domain is the long-term state drift that occurs when continuous-time models are exposed to prolonged seismic events or sequences of aftershocks. To address this issue, we introduce a learnable event-triggered discrete reset mechanism that monitors a prediction error signal and an auxiliary drift accumulation variable. When a trigger condition is met, the Liquid Neural Network hidden state is reinitialized to a learned baseline state, thereby preventing divergence from physically plausible structural dynamics. This reset mechanism is inspired by hybrid small-gain and impulsive control frameworks and provides rigorous stability guarantees. The continuous-time dynamics are modeled as a neural ordinary differential equation with a sinusoidal activation function, solved via a fixed-step Runge–Kutta integrator. The discrete-time pathway employs a two-layer LSTM with Bayesian-optimized hyperparameters. A learned attention mechanism fuses the hidden states from both pathways, and the combined representation is passed through a feedforward network to produce a four-class seismic vulnerability index. The entire system is trained end-to-end with an adaptive loss function that balances reconstruction accuracy, drift penalization, and classification performance. We first pre-train the model on synthetic data generated from a calibrated finite element model of a representative ancient hall in Rucheng, Hunan. Transfer learning then fine-tunes the model on real monitoring data. Our approach replaces conventional fragility-curve-based estimates with a data-driven, adaptive vulnerability assessment that is site-specific and robust to non-stationary excitation. The hybrid switched-system framework therefore offers a principled solution to the state drift problem while maintaining the expressive power of continuous-time neural dynamics.