Research on urban landscape big data information processing system based on ordinary differential equations

While big data brings quantitative analysis to landscape planning, it also introduces new analysis methods. In recent years, there are two common analysis methods for textual information, the analysis method for textual information, high-frequency word analysis and sentiment analysis. The high-frequency word analysis is similar to our common frequency distribution, but there are some differences. Frequency distribution is an effective analysis method, which divides the variable value into several intervals according to a certain interval value, and represents the distribution value of each interval[12]. The high-frequency word analysis is an analysis method that, for text data, counts the frequency of occurrence of each word, and sorts it according to the number of occurrences, so as to reflect the attention of the crowd. Although both methods search for regularities through frequency analysis, there are differences in the objects, methods, and scope of application of the two methods.

Sentiment analysis is similar to traditional psychological experimental survey analysis. Sentiment analysis is a software analysis of the text in travel notes or reviews, analyzing words with emotional color, and understanding tourists' evaluation and overall impression of scenic spots. The psychological experiment is to find the influence of the space environment on the psychology, the same method to quantify the psychological feeling. It can be seen from the definition that although the two analysis methods both study the user's feelings, attitudes, preferences and impressions, sentiment analysis is to collect and summarize the words expressing emotions without the user's knowledge., so as to analyze their emotional attitudes towards the research site. Psychological experiments are to quantify psychological feelings through questionnaires under the condition of the respondents' knowledge.

The garden information management function in the system mainly includes the management of landscape types in smart city gardens, the location query, greening cycle management and maintenance management. Garden information management is mainly realized through geographic information technology combined with big data technology. The system can more conveniently allow users to enter the relevant information query and management of smart city gardens through component geographic information technology. The layer vector data of information is introduced into the system, which is more convenient for the management of image information.

Managers manage various types of information such as garden landscapes by logging in to the system, perform related operations in the management function, and select the objects that need to be added for management. For example, green space, parks, street trees, etc. in the garden, so as to obtain the specific information of the corresponding garden. And then according to the search, edit and delete functions in the system to achieve its management operations. For the query of the garden location, you can query the specific garden location through the information list in the system management function, by selecting the object, city, street and other information to be queried. At the same time, the system also automatically refines the queried information, and marks key information such as the garden number, type, affiliated unit and maintenance personnel in the information, so as to facilitate user management. For the management of the landscaping cycle, the user selects the start and end time of landscaping to query the relevant specific information within a certain period of time. It includes information such as number information, name information, personnel information and maintenance status of the landscaping cycle. For garden maintenance management, the user selects different maintenance units and maintenance status in a maintenance cycle, and the system automatically queries the specific information within the corresponding range, and performs management operations such as querying and editing the information.

The three-dimensional display of smart city gardens should build a simulated garden display scene in the system according to the needs of the smart city. The specific operation steps of the construction are: first collect the corresponding data information, the information source of the display part mainly depends on the various garden data in the information management in the system, extract the garden information that needs to be displayed in the information management, and apply it to in building the scene. Secondly, the three-dimensional entity modeling of the buildings and landscapes in the garden is carried out, and the construction is carried out according to the collected garden data. When constructing, we should not only use 3D solid modeling technology, but also pay attention to the tree results in order to achieve the purpose of clear structure and provide sufficient conditions for the visual display in the system. Finally, scene integration uses the method of combining virtual and real to divide all kinds of buildings, plants, sky, and roads in the garden into different modules, and combine them according to certain rules according to the layout in the actual garden scene. In the display scene, the user can complete operations such as forward and backward in the virtual display scene through mouse operations. At the same time, the way of defining shortcut keys can also be used to realize automatic roaming in the virtual display scene of the system.

In order for the system to realize the real-time interaction capability between the information management and the 3D display scene, it is necessary to exchange information between the two every few frames, and then judge whether the information contains control information through the analysis of the two. If it exists, the two control the corresponding information respectively, so as to realize the real-time reception function of garden information in the information management of the three-dimensional display in the system. In order to have good interaction between the system and users, after completing the management and 3D display of garden information, it is necessary to synchronize the two to facilitate user operations.

The recommendation module is based on the WeChat public platform, taking WeChat users who pay attention to the public account of urban garden landscape as the primary target, and the friends of the target user as the secondary target. Focus on landscape-related WeChat users as a three-level target, and recommend urban garden landscape information to the above three-level target customers. The detailed recommendation algorithm is shown in formula (1): (1) Sab={1,if a connects with b and more than 1 node belongs to S0,else {S_{ab}} = \left\{ {\matrix{ {1,\,if\,a\,connects\,with\,b\,and\,more\,than\,1\,node\,belongs\,to\,S} \hfill \cr {0,\,else} \hfill \cr } } \right. In formula (1), S represents a community in the network, and a and b represent network nodes. The second and third-level target node screening methods are: (2) F=12×∑Sabγ(a,b)∑Sabδ(a,b) F = {1 \over 2} \times {{\sum {{S_{ab}}\gamma \left( {a,b} \right)} } \over {\sum {{S_{ab}}\delta \left( {a,b} \right)} }} In formula (2), F represents the degree of screening modularity. If nodes a and b belong to the S community, then (a, b) takes the value of 1. If the nodes a and b do not belong to the S community, then (a, b) takes the value of 0, if one of the a and b nodes belongs to the S community, then (a, b) takes the value 1; if the a and b nodes do not belong to the S community, then (a, b) takes the value 0.

The recommended steps are as follows: the first step, set the smart city garden landscape public account as a community; the second step, set the users who follow the public account as a node 1 set. As for the community, set the first target friends as node 2 set, the same as in the community; the third step, push urban garden information according to different levels. The fully convolutional neural network learning method is used for garden information feature classification and collaborative recommendation. Based on ordinary differential equations, the connection weights X (x1, x2, …, xn) from the hidden layer to the output layer are solved. An adaptive search is performed, and in the trust model of the fully convolutional neural network, the update rule for the collaborative recommendation of the fully convolutional neural network is expressed as an ordinary differential equation: (3) x˙k+1=xk−Ak−1gk {\dot x_{k + 1}} = {x_k} - A_k^{ - 1}{g_k} Based on the above analysis, the collaborative filtering method is used to train the connection weights of the hidden layer and the output layer to realize the collaborative filtering and classification recognition of the convolutional neural network, and the collaborative recommendation model is described as:

Input: According to the garden information test sample y to be recommended and the training sample set D, the connection weight distribution set of the nodes is obtained, a(y)=[1l,1l,B,1l]2 a\left( y \right) = {\left[ {{1 \over l},{1 \over l},\,B,\,{1 \over l}} \right]^2} , l is the width value of the basis function; output: the user interaction result α of the garden information collaborative recommendation.

According to the system designed in this paper, firstly select the ecological capacity, spatial capacity and tourism psychological capacity to analyze the environmental capacity of the system. The ecological capacity of various types of land use provided by the planning specification, the ecological capacity of the land use types in the demarcated site is shown in Table 1, and the space capacity and tourism capacity obtained through the topographic data can be obtained in Table 2.

It can be seen from Table 4 that among the three factors, the ecological capacity value is the smallest, which becomes the determining factor limiting the environmental capacity of the site. In the planning, the daily capacity of ecological capacity should be 54,800 people and the annual capacity should be 20,030,800 people as the red line for development.

The functions in the smart city garden information management and display system have been implemented one by one in the above, combined with the red line environment capacity analyzed in Section 4.1, in order to further verify the efficiency and interactivity of the system, the following will be the system and The traditional management system is compared with the experiment to further verify the operation efficiency of the system.

Randomly select 20 urban garden managers and 20 system users. The managers complete the operations of querying, editing, and deleting garden information, and the system users complete the operations of browsing, searching, and viewing the landscape information. Eliminate other factors that affect the experiment as much as possible, and complete the comparative experiment.

The relevant data information generated during the experimental operation is recorded, the result of testing the recommendation module is used as the experimental direction, and the user interaction rate after message push is used as the experimental index to verify the operation efficiency of the system in this paper. The interaction rate of the two systems is calculated, and the comparison chart of the experimental results shown in Figure 2 is drawn.

Figure 2

It can be clearly seen from the two curves in Figure 3 that through the repeated experiments of the experimenters, the user interaction rate when using the system in this paper to manage the information of smart city gardens shows an obvious upward trend, and the average user interaction rate is higher than 40%. When using the traditional system to manage the smart city garden information, the user interaction rate is significantly lower than the system in this paper, and it shows a downward trend. Therefore, this paper pushes messages for users of different levels in the recommendation module, and adopts the fully convolutional neural network learning method to classify garden information features and recommend collaboratively, thereby improving the user interaction rate. Therefore, it is further verified by experiments that the smart city garden information management and display user interaction rate designed in this paper is higher, the system operation efficiency is higher, and it is more in line with the actual management personnel in the garden information management process. User experience and comfort.