Jiangnan gardens have become a valuable cultural heritage of China with its elegant garden style. The article proposes a binocular visual recognition system by analyzing the composition of the garden spatial elements and performing feature fusion based on scene-driven coefficients. Ablation experiments are conducted on each part of the constructed data enhancement framework for generating the design of the Jiangnan garden plan, which is applied to generate a set of high-quality datasets and apply the data to image segmentation for generating the design of the Jiangnan garden. The algorithm training is carried out by applying the generated design plan dataset. On this basis, the data from the actual Jiangnan garden research and the spatially quantized feature data are used to do the correlation analysis between the design elements and the aesthetic mood. The data enhancement framework constructed in this paper improves the IOU of ST elements to 0.537, and the average intersection and merger ratio MIOU is 0.389. It shows that the data evaluation framework based on visual recognition is suitable for the study of plan generation of Jiangnan gardens. The correlation coefficients of connection value, spatial control value, average depth value, and integration degree regarding aesthetic context with the data of Jiangnan garden design elements are 0.173, 0.301, -0.278, and 0.325, respectively, which indicate that there is a significant correlation between all of them.

Based on the concept of “user-centered”, this paper designs a product form optimization model based on ant colony algorithm. Through mining the online reviews of the products, we determine the perceptual imagery of users, and categorize the perceptual imagery and determine the weights from the perspective of user satisfaction. Combining the factor analysis of perceptual imagery and the contribution value of morphological features on perceptual imagery, the product morphology optimization fitness function is constructed. Solve the model according to the basic principle of ant colony algorithm, and study the decision-making method to assist product optimization. Take a brand A model forum word-of-mouth data as an example to analyze, obtain users’ perceptual imagery through SO-PMI algorithm, and assign values to perceptual intention weights with the help of cluster analysis. Determine the contribution value of morphological features through the SD investigation of product morphological differences. Genetic algorithm is introduced to carry out comparative experiments to verify the superiority of ant colony algorithm in optimizing model solving. Finally, the application effect of the predictive model solving scheme is analyzed through user satisfaction survey. The results show that the output of the product optimization design model based on ACO algorithm Model A is 8. 23.11% of the users are very satisfied with the optimized Model A, 65.55% of the users are satisfied, and 85.72% of the survey respondents are very willing and ready to buy the optimized Model A.

The evaluation of English course goal attainment is an important basis for colleges and universities to judge whether the goal of cultivating foreign language talents has been achieved. This paper proposes a method for quantitative assessment of course goal attainment according to the OBE concept. Calculating the importance of attributes about classification, the decision tree algorithm based on rough set is proposed, combined with association rules for deep mining of educational data. Collect quantitative educational data and questionnaire data of a university, modeling relying on SPSS Modeler 14.2, and outputting decision tree of influencing factors. Using the evaluation of course goal achievement to analyze the achievement of A4 course goals, and exploring the association rules of influencing factors based on the decision tree. The traditional decision tree algorithm is introduced as a control group to evaluate the performance of the rough set-based decision tree algorithm. The results show that the achievement degree of each sub-objective of A4 course is higher than 0.70, and students who have the achievement degree of A4 course objective greater than 0.7, the nature of their major is foreign language and they have passed the Grade 4 test have a higher possibility of achieving the final foreign language talent cultivation goal of the university. The precision of the assessment method based on rough set decision tree is maintained at about 88%, and the accuracy rate is basically maintained at about 90%.

Frequent outbreaks of cyanobacterial blooms in Lake Taihu are undoubtedly a great threat to the economic development of its neighboring areas and the safety of drinking water of its residents. This paper takes Taihu Lake as the study area and analyzes its geographic location information and development status. Then, based on the remote sensing data from MODIS and Landsat 8 satellites, the normalized vegetation index is improved to identify the blooms, and the dynamic detection method of cyanobacterial blooms is constructed by combining with the remote sensing inversion of water temperature. At the same time, the spectral performance of each band is integrated to excavate the characteristic information of cyanobacterial bloom, and the algorithm in this paper is used to process the satellite remote sensing data of cyanobacterial bloom in Lake Taihu to analyze its spatial and temporal distribution characteristics, which is used as the basis of the dynamic warning model for early warning. Then the LightGBM method is introduced to realize the all-weather spatial and temporal continuous monitoring of cyanobacterial blooms in Lake Tai. Analyzing the monitoring data of this paper’s model on the intraday change process of cyanobacterial bloom in Lake Taihu, it is found that the trend of intraday change in the area of cyanobacterial bloom in Lake Taihu in different seasons is relatively consistent, with the highest area of the bloom in autumn, accounting for 21% of the area of Lake Taihu’s water body. The study pointed out that after entering the fall, extra attention should be paid to the monitoring, prevention and control of cyanobacterial bloom in Lake Taihu.

In this paper, we design and implement a model network for English writing style generation using UNet network as well as ViT for encoding and decoding, and PatchGAN to enhance the identification speed. Based on the CRF-NLG model to identify and extract professional English terms, and design a special loss function to optimize the quality of writing style generation. The F1 value is used to evaluate the model recognition ability, and the writing style generation effect is explored by controlled experiments of the proposed model and three baseline models. The practical application results of the proposed model are visualized from four perspectives: overall evaluation, style strength, content preservation, and fluency, to verify its practical application effect. The results show that the proposed model exhibits the strongest performance in the two levels of content preservation and fluency, which are improved by 12.71% and 39.11%, respectively, compared with the existing GAN-based style generation model. Of the 119 modifications 92 (77.3%) were better, 17 (14.3%) were average, and only 11 (9.2%) were worse.

With the rapid development of science and technology, the traditional mode of teaching is inefficient and difficult to flexibly respond to the needs of knowledge updating, and generating content and applications based on AI has become an important way to solve this problem. According to the form of interaction in the digital exhibition hall, the article proposes SinGAN model and uses the multi-head self-attention mechanism to coordinate the overall features and detailed features in the generated adversarial network image, and to deal with the large range of dependencies in the image. The proposed AI-generated content and SinGAN image processing method are applied in the teaching of practical courses using the course “Digital Electronics Technology and Application” of a university in Guangdong Province, which specializes in electronic information and engineering, as an experimental object. The experiment shows that the percentage of content with a content quality score of 0.6 to 1.0 reaches 75.7%. As the course progresses, the keyword coverage rate reaches 0.996, and AI-generated content is efficiently applied in the course. The student performance of the experimental class with AI-generated content and image processing method teaching mode and the regular class with traditional teaching mode were 80.75 and 67.91 respectively, and the sample t-test for the significance of the student performance of the two classes was P=0.006, which showed a significant difference in the students’ performance between the two teaching modes. Students’ satisfaction with the new teaching mode is high, indicating that the AI-generated content and image processing methods proposed in the article have been well applied in education reform.

With the arrival of the big data era, the demand for massive data storage is growing, and distributed storage systems have become a key technology to solve this problem. The traditional HDFS system has a large storage overhead, this paper in order to improve the storage efficiency of massive data, the introduction of corrective deletion code (RS code) technology, to ensure the reliability of the data at the same time significantly reduce the cost of storage. In order to improve the storage efficiency of massive data, this paper introduces the corrective censoring code (RS code) technology, which ensures the data reliability and significantly reduces the storage cost. In addition, to address the problems of low coding efficiency and high repair overhead in the practical application of RS code, this paper further introduces the local repair code (LRC) technology, which reduces the data repair overhead, and compares and analyzes the application effect of optimization model (RS-LRC-HDFS). The experimental results show that after RS-LRC optimization, the time overhead of the HDFS storage system in the write process and read process is significantly improved by 81.12% and 93.01%, respectively, compared with the pre-optimization period, and the repair time of massive file data is reduced by 87.25%. It can be seen that it provides an efficient and reliable solution for massive data storage.

This paper intends to introduce the multi-intelligence of digital resources in cultural and tourism industry in reinforcement learning. In order to scientifically evaluate digital resource allocation, the index system characterizing resource allocation is constructed using hierarchical analysis. From there, a multi-objective collaborative optimization allocation model of digital resources in cultural and tourism industry based on reinforcement learning and multi-intelligent body system is established. Through empirical analysis, it can be seen that referring to the observation of the development of the comprehensive level of digital resource allocation, there is an imbalance in the development level of N province. The indicator system is refined to consist of 4 guideline level indicators and 26 indicator level indicators. Before and after the multi-objective synergistic optimization, the total amount of digital resource procurement for the cultural and tourism industry in province N was reduced by 460,742 yuan. After optimization, the comprehensive efficiency of resource allocation in area a improves by 0.03136, area b improves by 0.03275, and area h improves by 0.02799. Moreover, all of them tend to be in equilibrium. Therefore, the multi-objective synergistic optimization allocation model in this paper can improve the efficiency of digital resources in cultural tourism industry and reduce the differences between districts and counties.

Based on the Delphi method and relevant definitions, this paper determines the evaluation index system of college students’ employability, adopts the hierarchical analysis algorithm (AHP) to calculate the weights of the evaluation indexes, and for the weights of the evaluation indexes do not satisfy the consistency test, adopts the Adaptive Gradient Algorithm (AdaGrad) to adjust the weight parameters so as to make them satisfy the consistency test, and arrives at the adjusted values of the evaluation indexes weights. The weights of the adjusted evaluation indexes are derived. Using the fuzzy comprehensive evaluation theory, a comprehensive assessment model of college students’ completion ability was constructed, and then the research sample was evaluated and analyzed with the help of this model. It is calculated that the affiliation vector of the evaluation of college students’ employability is (1.9466, 1.2539, 1.1123, 0.9752, 4.714), and the maximum affiliation value is 4.714, which can be inferred that the students of this university have good comprehensive ability of employment and can well face the employment pressure in the current society.

Micro-landscape is a kind of green landscape designed to enhance the local landscape environment of the city along with the renewal of urban green space and the transformation of old city. The article adopts Hadoop technology and utilizes the Hadoop distributed computing framework to preprocess the data, constructs the urban micro-landscape greening evaluation system, and carries out research on four evaluation levels, namely, building façade landscape, multimedia landscape, water landscape, and landscape facilities. At the same time, based on the principal component analysis and factor analysis method for comprehensive evaluation, it is determined that the multimedia interaction factor is the most important factor affecting the effectiveness of micro-landscape greening. Then use SWMM model to design a city urban area, through SWMM model simulation to get the actual average annual runoff control rate of the demonstration area in 2023 is 59%, and the overall long-term goal of urban micro-landscape greening planning in 2020-2030 there is a gap, based on which put forward the urban micro-landscape greening design program.