16 results about "Spatial classification" patented technology

The invention relates to the field of visual processing, and provides a visual classification processing method and device based on large model and multi-modal data fusion. The method comprises the following steps: inputting a to-be-classified input image and a corresponding category text description into a text encoder for multi-level feature extraction to obtain global text features and local text features; performing fine-grained cross-modal alignment on the local visual features and the local text features, calculating association weights between the image regions and the text phrases through a bidirectional cross attention mechanism, and generating aligned intermediate features; splicing and fusing the aligned middle features and the global visual features, and inhibiting background noise in a fusion result and reinforcing discriminative features in the fusion result through a feature mask algorithm in combination with the global text features to obtain multi-modal fusion features; and synchronously inputting the multi-modal fusion features into a multi-space classifier to generate respective classification results, and adaptively outputting an image classification result according to a confidence threshold in combination with a dynamic routing mechanism.

The invention discloses a cultivated land ecological safety comprehensive evaluation method and system based on multi-model coupling. The method comprises the steps that a future multi-source data set of cultivated land to be evaluated in a future time period is set; performing grid division on the cultivated land to be evaluated by adopting a grid calculator in ArcGIS software to obtain a plurality of grid regions, and performing grid division on the future multi-source data set based on the plurality of grid regions to obtain a future multi-source sub-data set of each grid region; and predicting the future multi-source sub-data set of each grid region by adopting the constructed multi-ecological safety comprehensive evaluation function model, the spatial classification method and the constructed ecological safety comprehensive evaluation index calculation formula to obtain an ecological safety comprehensive evaluation index corresponding to each grid region. According to the method, the ecological function of the to-be-evaluated cultivated land in the future scene is simulated by coupling the multi-ecological safety comprehensive evaluation function quantity model, so that systematic evaluation on the cultivated land ecological function of the to-be-evaluated cultivated land in the future scene is realized.

The application discloses a kind of based on the cultivated land ecological safety comprehensive evaluation method and system of multi-model coupling, which comprises: setting the future multi-source data set of the cultivated land to be evaluated in future time period;Using the grid calculator in ArcGIS software, the grid division is carried out on the cultivated land to be evaluated, a plurality of grid regions are obtained, and the future multi-source data set is divided into grid based on the plurality of grid regions, to obtain the future multi-source sub-data set of each grid region;Using the multi-ecological safety comprehensive evaluation function quantity model that has been constructed, spatial classification method and the ecological safety comprehensive evaluation index calculation formula that is constructed, the future multi-source sub-data set of each grid region is predicted, to obtain the ecological safety comprehensive evaluation index corresponding to each grid region.The application simulates the ecological function of the cultivated land to be evaluated under future scenario by coupling multi-ecological safety comprehensive evaluation function quantity model, to realize the system evaluation of cultivated land ecological function of the cultivated land to be evaluated under future scenario.

The application discloses a landslide susceptibility comprehensive evaluation method, system, device, medium and product, relates to the technical field of geological disaster prediction, and the method comprises the steps: taking the SBAS-InSAR deformation rate as a dynamic deformation factor, combining a static environmental factor, and constructing an original feature dataset; adopting a recursive feature elimination algorithm, performing importance sorting and iterative elimination on evaluation factors in the original feature dataset, determining an optimal feature subset, training an RFE-RF-XGBoost cascade integrated model; inputting the optimal feature subset corresponding to all grid units in the research area into the trained RFE-RF-XGBoost cascade integrated model, determining a landslide susceptibility spatial classification map, verifying and locally correcting different risk level regions divided in the landslide susceptibility spatial classification map, and determining a landslide susceptibility evaluation result, so that the timeliness and accuracy of the landslide susceptibility evaluation result are improved.

The invention discloses a carbon emission intensity prediction simulation and land utilization management and control method and system, and belongs to the technical field of carbon emission calculation prediction.The method comprises the steps that firstly, multi-source space and statistical data are integrated to establish a basic database, and the carbon emission intensity of all kinds of space is calculated according to land space classification; then, constructing an extended STIRPAT model to screen a core driving factor, and constructing a feature project containing residual, lag and trend features based on the output of the extended STIRPAT model; training an XGBoost machine learning model by adopting a Bayesian optimization algorithm so as to capture a complex nonlinear relationship between the carbon emission intensity and the influence factors; and finally, inputting future scene feature data, generating a territorial space carbon emission intensity distribution map in combination with a PLUS model, and identifying a high-emission hot spot region. According to the method, the problem that the technical progress influence is difficult to quantify and the spatial explicit prediction is difficult to carry out in the prior art is solved, the prediction precision is remarkably improved, and a scientific basis and decision support are provided for regional territorial space planning and low-carbon management and control.

The invention discloses a campus carbon emission prediction and dynamic regulation and control method. The method comprises the steps of data acquisition and preprocessing, feature engineering, static modeling, time sequence prediction, dynamic regulation and control and feedback optimization. The method comprises the following steps: collecting multi-source heterogeneous data such as campus personnel density, equipment use, space classification and meteorology, constructing a regional carbon emission intensity function by using an XGBoost regression model, and capturing a static nonlinear relationship; and learning a historical carbon emission sequence by using an LSTM network, predicting the carbon emission trend in the next 24 hours, and identifying peaks and anomalies. And on the basis of comparison between the prediction result and a threshold value, differential regulation and control strategies for the teaching area, the dormitory area and the canteen area are generated, and equipment control and personnel behavior guidance are executed through the BAS system and the mobile terminal. On-line learning is carried out by using actually measured data and prediction deviation, model parameters are optimized, and a prediction-regulation-feedback closed loop is formed. According to the method, aiming at campus functional area heterogeneity and personnel flow characteristics, high-precision prediction and active intervention are realized, the carbon emission peak value is reduced, the energy efficiency is improved, and sustainable development is supported.

This invention relates to the field of urban planning technology, and more particularly to a historical and cultural-oriented spatial classification and regeneration method and system. The method includes the following steps: acquiring historical and cultural data, and performing multimodal perception on the historical and cultural data to obtain historical and cultural spatial data; extracting deep features from the historical and cultural spatial data to obtain historical and cultural element data; constructing a spatial semantic map from the historical and cultural element data to obtain spatial semantic map data; classifying the spatial semantic map data according to its chronological evolution trajectory to obtain historical and cultural spatial classification data; generating spatial regeneration data based on the historical and cultural spatial classification data to obtain spatial regeneration data; acquiring urban planning demand data, and integrating the urban planning demand data and the spatial regeneration data to obtain integrated spatial regeneration data. This invention combines urban planning with spatial regeneration, effectively achieving a dynamic balance between historical and cultural preservation and modern urban development.

The invention provides an urban rail transit intersection optimization method and device and electronic equipment, and the method comprises the steps: employing a K-means clustering algorithm to carry out the spatial classification of construction environment factors of rail transit line stations, and recognizing the types of stations with different functional features; based on the space classification result of the built environment, a random forest algorithm is adopted to construct a passenger flow prediction model fusing multi-dimensional built environment elements, and the passenger flow volume of the rail transit stations along the line is predicted to obtain a passenger flow prediction result; based on the space classification result and the passenger flow prediction result, constructing a large and small intersection multi-target optimization model fusing the built-up environment factors; and solving the large and small intersection multi-target optimization model by combining a multi-target genetic algorithm to obtain a target optimization scheme of the rail transit. According to the invention, a conversion mechanism from space cognition to operation decision is established, and the problem of poor route optimization effect of urban rail transit in the prior art is solved.

This invention belongs to the field of data processing and discloses a deep learning-based method for dynamic monitoring of blue-green space in urban renewal. The method includes: acquiring high-resolution remote sensing images and preprocessing them to obtain preprocessed image data; determining a UNetMamba model and improving it using a ResTv2 backbone network and ECA attention mechanism to obtain an improved UNetMamba model; determining a preset dataset and training and optimizing the improved UNetMamba model based on the preset dataset to obtain a blue-green space classification model; inputting the preprocessed image data into the blue-green space classification model and outputting the blue-green space classification results; and performing post-processing and spatial transfer econometric analysis on the blue-green space classification results to obtain a blue-green space distribution map and a blue-green space transformation Sankey diagram. This invention can effectively improve the identification accuracy of blue-green space, thereby effectively revealing the evolutionary patterns of ecological space.

The present specification relates to the field of oil and gas field exploration and development, and relates to a position coding determination method and a geological feature prediction method based on a well logging evaluation large model, the method comprising: identifying a to-be-coded geographical region according to a preset spatial classification rule, and determining a spatial range type to which the to-be-coded geographical region belongs; if the spatial range type is a first spatial range, determining a longitude weight coefficient and a latitude weight coefficient in the first spatial range according to the maximum longitude and latitude and the minimum longitude and latitude in the first spatial range, wherein the first spatial range is a defined spatial range; and calculating the position coding of each point in the first spatial range according to the longitude weight coefficient and the latitude weight coefficient. The present scheme comprehensively considers the elements of geological statistics, and position codes the position information of space, so that the large model based on the oil and gas industry can better consider the spatial correlation of geological information during pre-training, and effectively improve the precision of the oil and gas large model.

The invention relates to the technical field of data processing, and particularly discloses an AI-based air quality prediction method and system, and the method comprises the steps: S1, dividing a preset region into independent grids of a preset size, and setting the information of each grid; s2, extracting influence factors of grid air quality, wherein the influence factors comprise influence factors of spatial change and influence factors of time change; s3, constructing a space classifier based on an artificial neural network algorithm, and capturing the influence of space factors on the air quality; constructing a time classifier based on a convolution LSTM algorithm, and learning time periodic features in the time sequence data; and S4, through a collaborative algorithm, according to results of the space classifier and the time classifier, obtaining gridding air quality data with a time sequence. By adopting the technical scheme of the invention, the air quality prediction accuracy of mountainous regions and river valley cities can be effectively improved.

The invention relates to the technical field of data processing, and discloses a touch operation accurate identification method and system of a smart bracelet, and the core of the method is to collect capacitance, motion sensing and pressure multi-dimensional data of a bracelet touch screen, establish an analysis time window, extract features of each dimension and complete timestamp alignment and normalization. Constructing a multi-modal feature vector; a time sequence causal judgment result is output through causal test, a joint judgment space is constructed to obtain a space classification judgment result, and a dynamic threshold judgment result is output in combination with a motion intensity dynamic threshold value; and finally, integrating three types of results based on a Bayesian decision theory, outputting a final recognition result, collecting feedback data to optimize model parameters, updating a sample library, and forming a closed-loop recognition mechanism. According to the method, true and false touch control is effectively defined through multi-modal data processing and multi-dimensional judgment, the recognition precision is continuously improved through closed-loop optimization, misoperation is reduced, and the reliability of the intelligent bracelet touch control system is greatly enhanced.

This invention discloses a blue-green infrastructure mapping and classification method and related apparatus, relating to the field of computer technology. The method includes: generating a new reservoir layer and a first residual water body layer based on original water body layer data, water patch layer data, and point-of-interest layer data; generating a new river layer and a second residual water body layer based on the first residual water body layer and water patch layer data; generating a new lake layer and a new pond layer based on the second residual water body layer and multi-element patch layer data; generating a blue-green space layer based on land cover raster layer data and the above layers; performing land cover type area ratio analysis based on the blue-green space layer using field statistics; performing primary classification using the BGI classification framework based on the land cover type area ratio information; and performing blue-green space classification based on the primary classification results using the BGI classification framework. This invention improves the completeness and feasibility of BGI mapping and classification.

The application provides a mining area disturbance space identification method, system, computer device and medium, and belongs to the image recognition field, comprising: acquiring remote sensing images of a mining area; performing feature extraction on the remote sensing images to obtain a plurality of coding feature maps of different scales; performing weighted fusion on the plurality of coding feature maps of different scales to obtain a multi-scale fusion feature map; calculating edge gradients of the remote sensing images and the multi-scale fusion feature map respectively, and fusing to obtain a normalized edge feature map; determining an edge attention weight according to the normalized edge feature map, and performing weighted fusion on the edge attention weight and the coding feature map of the corresponding scale to obtain an edge guided feature map; and performing resolution reconstruction on the edge guided feature map to obtain a disturbance space classification image of the mining area. The method improves the accuracy of the mining area disturbance space identification, and meets the fine identification requirement of the disturbance space.

Examples of generating a map for an indoor space are described. In one example, a raw digital representation of the indoor space is obtained and preprocessed by at least one of marking a region-of-interest (ROI) in the raw digital representation and classifying the indoor space into a space-category to obtain a preprocessed representation. Using the preprocessed representation, a skeletal map is generated including indicators to identify elements of the indoor space. Revision inputs are received to modify an erroneous indicator from amongst the multiple indicators. The erroneous indicator is revised to a modified indicator and the modified indicator and unmodified indicators are used to create a modified skeletal map which is usable for localizing a mobile device in the indoor space.

The invention discloses a blue-green infrastructure drawing classification method and a related device, and relates to the technical field of computers, and the method comprises the steps: generating a new reservoir map layer and a first residual water map layer based on raw water map layer data, water plaque map layer data and interest point map layer data; generating a new river layer and a second residual water body layer based on data of the first residual water body layer and the water body plaque layer; generating a new lake map layer and a new pit-pond map layer based on the second residual water body map layer and the multi-element plaque map layer data, and generating a blue-green space map layer based on the land coverage grid map layer data and the map layers above; carrying out land coverage type area proportion analysis by utilizing field statistics based on the blue-green space layer; and performing first-level classification by using a BGI classification framework based on the land coverage type area proportion information, and performing blue-green space classification by using the BGI classification framework based on a first-level classification result. According to the method, the integrity and feasibility of BGI drawing classification are improved.