44 results about "Spatial code" patented technology

The invention discloses a city-level CIM scene streaming rendering method based on multi-modal fusion, and the method comprises the following steps: S1, collecting city data, and carrying out the preprocessing of the city data; s2, establishing a cross-modal entity alignment relationship, and constructing a city object structure body; s3, dividing a space by adopting a quadtree method, generating rendering sub-blocks and distributing a unique space code; s4, binding the rendering sub-blocks with the urban object structure body, constructing a multi-modal fusion network, and generating a multi-modal rendering unit set; s5, constructing an inter-frame rendering cost function, and scheduling and sequencing the multi-modal rendering unit set to form an inter-frame scheduling sequence; s6, loading the rendering units according to the inter-frame scheduling sequence to form a city-level three-dimensional view of continuous frames; and S7, updating the scheduling sequence, dynamically adjusting the loading sequence, and realizing stable and efficient streaming rendering of the city-level CIM scene. According to the method, the fusion efficiency and rendering performance of the urban large-scale three-dimensional data are improved, and the method has remarkable engineering applicability and popularization value.

The invention relates to a method for constructing a space-time multi-task prediction large model based on data fusion. The method comprises the steps that firstly, a data fusion module is constructed, the data fusion module comprises a time sequence alignment sub-module, a region fusion sub-module and a feature splicing sub-module, and original data sets are fused based on the three dimensions of time, space and features; then, constructing an input embedding module, and extracting time features of the data set while embedding the fused data set into a high-dimensional space; constructing a spatial coding module, and extracting spatial features of the data set; then, a base large model prediction module is constructed, and reasoning prediction is carried out; and finally, designing a multi-task output module, and performing multi-task splitting on a prediction result to obtain a space-time multi-task prediction large model. According to the method, multiple tasks can be predicted at the same time, the multi-task prediction capability of the model is improved, the calculation cost and deployment cost are reduced, and the method has great advantages in an urban traffic comprehensive scene needing quick response.

The invention relates to the technical field of intelligent traffic, in particular to an intelligent scheduling method and device for unmanned taxis, and the method comprises the steps: constructing a spatial topological structure for connecting a target unmanned taxi, passengers and a charging station based on the received vehicle data, passenger data and environment data; capturing a spatial dependency relationship among the target unmanned taxi node, the passenger node and the charging station node to generate a spatial coding feature and a time coding feature, and combining the spatial coding feature and the time coding feature with corresponding weights to generate a spatial-temporal feature of the target unmanned taxi; and generating a passenger matching result and / or a path plan and / or a charging decision of the target unmanned taxi based on the spatial-temporal characteristics. Therefore, the problems that a traditional unmanned taxi scheduling method does not fully consider complex factors in an actual driving scene, scheduling decisions are still mostly single-target guiding, dynamic adaptability is lacked, and the increasing travel requirements of passengers cannot be met are solved.

The embodiment of the invention discloses a self-supervised emotion recognition method based on electroencephalogram signals in a virtual reality scene. One specific embodiment of the method comprises the following steps: collecting original full-channel electroencephalogram data, and generating preprocessed full-channel electroencephalogram data; obtaining each differential entropy feature; generating each embedded feature; obtaining a time visible frame feature and a time masking frame feature; obtaining each piece of time code embedded information; obtaining each real two-dimensional coordinate, and generating each spatial position code; obtaining each space visible frame feature and each space masking frame feature; each piece of space code embedding information is obtained; obtaining each space decoding feature; obtaining reconstructed differential entropy features; determining a loss value, and determining a pre-training model; obtaining an emotion predictor; and obtaining an emotion category. According to the embodiment, the emotion recognition capability of the few-channel electroencephalogram data can be enhanced, and the problem that the emotion recognition accuracy is low due to individual differences can be solved.

The invention discloses a voice-driven speaking portrait generation method and electronic equipment. The method comprises the following steps: acquiring an audio feature, a spatial feature and an image feature; inputting the audio features and the spatial features into an audio-spatial coding module, and aligning the audio features with the spatial features by dynamically adjusting the spatial positions of the audio features and the face region to obtain spatially perceived lip features; inputting the audio features and the image features into an audio-guided dynamic fusion module, and calculating audio-perceived channel attention and space attention by using a dynamic linear layer to obtain dynamic fusion features; obtaining reference key points, inputting the reference key points and the audio features into an expression optimization module, and fusing the reference key points and the audio reference points to obtain key point features; and inputting the lip sensing feature, the dynamic fusion feature and the key point feature into a trained generative network to generate a video frame. According to the method, the higher-quality, more natural and stable voice-driven portrait synthesis effect is achieved.

The embodiment of the invention discloses an AI generative model training method, an image generation method and electronic equipment, and the method comprises the steps: obtaining a plurality of training data sets used for training an artificial intelligence AI generative model, and the plurality of training data sets comprise a plurality of condition images corresponding to a plurality of different control types; respectively carrying out feature extraction on the condition image and the text content of the control instruction through an adapter network of an AI generative model so as to obtain control information; and injecting the control information into a backbone network of the AI generative model in a cross attention mode, introducing the spatial coding information of the condition image into the backbone network, and completing training of the AI generative model. Therefore, the AI generative model can be uniformly controlled through the condition images of various different control types. According to the embodiment of the invention, unified control on the model by various different types of control signals can be realized.

The invention relates to a unified coding method and device of geoscience data, computer equipment and a storage medium. The method comprises the following steps: acquiring multi-modal geoscience data, wherein the multi-modal geoscience data comprises spatial position information and lithosphere information; setting the grid precision of the two-dimensional map grid, and performing mapping coding on the spatial position information according to the grid precision to obtain a spatial code of the spatial position information in the two-dimensional map grid; constructing a rock stratum knowledge base, wherein the rock stratum knowledge base comprises a plurality of stratum unit terms with different priorities; and according to a plurality of stratigraphic unit terms with different priorities, extracting a target field in the rock stratum information, and carrying out Hash coding on geoscience data stored under the target field to obtain a group data code of the rock stratum information. By adopting the method, the unified coding of the multi-modal geoscience data can be realized by carrying out spatial coding on the spatial position information and carrying out group coding on the rock stratum information.

The invention discloses a real-scene three-dimensional system geographic entity space-time coding method. According to the method, the dynamic sensing capability of the geographic entity state is remarkably improved through an evolution prediction mechanism. Traditional coding mostly focuses on current state records, but the method depends on a state transition probability model of time decay weighting, can accurately capture recent change trends of entities, generate evolution prediction codes and dynamically bind the evolution prediction codes with space codes and time codes, so that the coding can not only reflect current positions, time and attributes of the entities, but also can reflect current state records of the entities. And the state trend of 1-3 months in the future can be pre-expressed. According to the method, the subsequent stage of a construction road can be pre-judged in advance, an emergency department can quickly recognize potential risk entities, the problem that geographic information lags behind actual change is effectively solved, more prospective data support is provided for collaborative decision making of multiple departments, the suitability and long-term reliability of coding in multiple scenes are enhanced through full-process closed-loop design, and the method is suitable for popularization and application. And a stable basic support is provided for efficient application of live-action three-dimensional geographic information.

The invention provides a time domain mapping encryption imaging method and system based on an event camera, and belongs to the field of computer vision hardware encryption imaging. The method comprises the steps that a receiver builds an asynchronous time domain mapping imaging system and constructs a space coding secret key sequence, then obtains an SLM modulation pattern sequence, obtains a noise SLM modulation pattern sequence and sends the SLM modulation pattern sequence to a sender; and the receiver carries out calibration based on the SLM modulation pattern sequence to obtain a space decoding secret key sequence. And the sender obtains the imaging event information and the encrypted time decoding secret key sequence based on the noise SLM modulation pattern sequence and sends the imaging event information and the encrypted time decoding secret key sequence to the receiver. And the receiver obtains a fragmented grey-scale map based on the spatial decoding secret key sequence and the received information, and then obtains a dense and complete grey-scale image through a video completion network based on the fragmented grey-scale map. According to the invention, real-time encryption can be realized, and the leakage risk before encrypted information transmission is reduced; moreover, the method is not sensitive to the collection environment, is high in decoding efficiency, and is low in implementation difficulty.

The invention discloses a space coding and decoding method for steady-state visual evoked potential, which realizes mapping and identification of a target and a space characteristic by combining a multi-frequency characteristic with a canonical correlation analysis method and utilizing a space coding characteristic of a stimulation target. By monitoring and analyzing SSVEP response distribution of surrounding targets, different central targets can be effectively distinguished, so that spatial coding is realized. In the decoding process, the target frequency corresponding to the most relevant frequency of the electroencephalogram signal is identified through the maximum correlation coefficient. Then, in combination with the frequency of the surrounding target, an inclusion index is calculated to further confirm the validity of the identification. And when the inclusion degree is greater than a set threshold value, considering that the current center target is identified as an effective center target. According to the method, the recognition precision can be improved, the electroencephalogram channel requirement is reduced, the information transmission rate and the real-time performance of the system are improved through the optimized space coding scheme, and the method is particularly suitable for low-delay and multi-target recognition brain-computer interaction tasks.

Embodiments feature families of rate allocation and rate control methods that utilize advanced processing of past and future frame / field picture statistics and are designed to operate with one or more coding passes. At least two method families include: a family of methods for a rate allocation with picture look-ahead; and a family of methods for average bit rate (ABR) control methods. At least two other methods for each method family are described. For the first family of methods, some methods may involve intra rate control. For the second family of methods, some methods may involve high complexity ABR control and / or low complexity ABR control. These and other embodiments can involve any of the following: spatial coding parameter adaptation, coding prediction, complexity processing, complexity estimation, complexity filtering, bit rate considerations, quality considerations, coding parameter allocation, and / or hierarchical prediction structures, among others.

The invention discloses a method for establishing an entity object relationship of urban governance elements. The method comprises the following steps of: firstly, dividing urban governance multi-source data into spatial data and non-spatial data; giving spatial codes to the spatial data according to a spatial grid where a central point of the spatial data is located; especially for the planar elements, a tolerance calculation mechanism is introduced, when spatial overlay analysis is carried out, only when the geometric overlapping relation between the planar elements and other spatial elements meets the preset tolerance calculation condition, it is judged that the planar elements have the inclusion relation and correlation is established, and therefore boundary overflow errors caused by surveying and mapping precision are corrected. And for non-spatial data, accurately mounting the non-spatial data to the established spatial entity through address standardization matching of place name address objects or keyword analysis. And finally, on the basis of a unique space code formed by combining the professional identification code, the space identification code and the time identification code, constructing an urban governance element entity object. According to the method, the problems of inaccurate spatial logic judgment and difficult association of non-spatial data in traditional data fusion are effectively solved, and accurate fusion and efficient indexing of urban governance data are realized.

Consistent with an aspect of the present disclosure, an exemplary digital communication system is provided that includes redundant or spare transceivers to handle failure events. In one example, if one of the first N+r transceivers experience a failure, spatial FEC codewords can correct the resulting bit errors and, as previously mentioned, prevent post-FEC errors from occurring. In embodiments, if a primary transceiver fails, a redundant transceiver may be activated to take its place.

Various aspects of the present disclosure generally relate to wireless communication. In some aspects, a user equipment (UE) may receive one or more channel state information (CSI) reference signals (CSI-RSs). The UE may select a first spatial domain (SD) basis from a spatial codebook based at least in part on a measurement of the one or more CSI-RSs. The UE may select remaining SD bases within a same orthogonal oversampling group as the first SD basis, where a size of the orthogonal oversampling group is N1 by N2 of the spatial codebook, and N1 and N2 are dimensions representing a quantity of CSI-RS ports. The UE may transmit an indication of the selected first SD basis and the selected remaining SD bases in a CSI report. Numerous other aspects are described.

The invention relates to the technical field of aviation safety monitoring, provides a flight personnel operation state identification method and system based on multi-modal data fusion, and solves the problems of high misjudgment rate and poor environmental adaptability in the prior art. The method comprises the following steps: synchronously acquiring electroencephalogram signals of a flight personnel, angular velocity of a head inertial measurement unit, an eye movement track and pitch angle and roll angle parameters of an aircraft; dynamically segmenting the electroencephalogram signal to construct an electroencephalogram dynamic topological graph, and generating a head stability index; extracting wave persistence features of the forehead lobe brain region, and performing weighted fusion on the wave persistence features and the head stability index to generate a cognitive conflict feature vector; the eye movement track coordinates are converted into a cabin three-dimensional coordinate system to generate space codes, and space-time feature bodies are constructed through space-time convolution kernel fusion space distribution and track density gradients; and identifying the operation state to be normal, early-warning or out-of-control three levels. According to the invention, accurate grading early warning of the operation state of the flight personnel is realized, and the real-time performance and the accuracy of flight safety monitoring are remarkably improved.

The invention discloses a land full-process approval management method, particularly relates to the technical field of land intelligent management, and realizes full-period run-through, cross-system collaboration and intelligent decision by constructing a dynamic space code mechanism. A land parcel is taken as a core, a static sequence (GeoSOT grid code) is adopted to guarantee spatial position consistency, a dynamic sequence (business state code / timestamp / serial number) maps an examination and approval state in real time, and code updating and associated business synchronization are automatically triggered when the land parcel is changed; based on a national project code and a space code series tax, municipal and financial department system, external data is called through a government affair data sharing platform, and one application and full sharing are realized; spatial conflicts between land and permanent basic farmland and ecological red lines are automatically identified and early warned by utilizing GIS topological overlay analysis, a full-process structured data link (WKT / JSON format) is constructed in combination with fuzzy retrieval, the problems of stage splitting and data islands are solved, and approval accuracy and collaborative efficiency are improved.

The present invention provides a method for constructing a target association tracking dataset based on self-reported position data. Scenario parameters are set based on the sensor sources of the multi-source target tracks to be constructed. The desired tracks are extracted from publicly available civilian AIS data, and the target density and maneuverability within the spatial grid are statistically analyzed. This data is used as a basic database for track simulation and stored in CSV format as an AIS spatial coding index file. Based on the target density and maneuverability within the spatial grid, AIS track data for the desired grid is extracted from the AIS spatial coding index file. Based on the extracted real tracks, a multi-sensor target track dataset that meets the set scenario parameters is constructed through coordinate translation, random batching, systematic error generation, random error generation, and track association mapping. This method constructs a unified, standardized, large-scale, high-quality dataset based on publicly available data, providing a fundamental guarantee for training deep learning neural networks for multi-source track association.

The embodiment of the invention relates to the technical field of three-dimensional graphic rendering, and discloses a mixed rendering method based on voxel coding. According to the method, for a three-dimensional scene in which multiple rendering strategies coexist, voxel-level division is carried out through the scene, space coding of nodes is calculated based on finite element grids, and ordered arrangement and linear coding of the nodes in the three-dimensional space are achieved. And by using the codes, constructing a code-driven scene block scheduling mechanism, and binding corresponding rendering strategies (such as neural rendering, point cloud rendering and traditional graphic rendering) to different coding areas. In the rendering stage, a proper rendering mode can be selected according to observation positions, equipment capabilities or application requirements, and unified management and mixed scheduling of multiple rendering modes are realized by means of a unified spatial coding index. The method is suitable for meta-universe construction, digital twinborn visualization and scene organization in a complex three-dimensional graphic system, and the scheduling efficiency and space consistency in a large-scale environment are improved.

The application discloses a speech-driven talking portrait generation method and electronic equipment, wherein the method comprises the following steps: acquiring audio features, spatial features and image features; inputting the audio features and the spatial features into an audio-spatial coding module, aligning the audio features and the spatial features by dynamically adjusting the spatial position of the audio features and the facial region, and obtaining spatially perceived lip features; inputting the audio features and the image features into an audio-guided dynamic fusion module, calculating audio-perceived channel attention and spatial attention by using a dynamic linear layer, and obtaining dynamic fusion features; acquiring reference key points, inputting the reference key points and the audio features into an expression optimization module, fusing the reference key points and the audio reference points, and obtaining key point features; and inputting the lip perception features, the dynamic fusion features and the key point features into a trained generation network to generate video frames. The method realizes a higher-quality, more natural and stable speech-driven portrait synthesis effect.

The application discloses a depression evaluation method based on audio-visual multi-modal data fusion, comprising the following steps: step 1, obtaining the facial video of a subject and the audio of the subject; step 2, obtaining the low-level visual features and low-level audio features of the subject; step 3, inputting the low-level visual features and low-level audio features obtained in step 2 into a parallel multi-scale bridge fusion depression evaluation PMBFN network for processing and obtaining a depression rating; by constructing a spatial coding module of a visual and audio branch, a parallel multi-scale dynamic convolution module and a space-time attention pooling module, multi-scale deep features are quickly and efficiently extracted from audio-visual multi-modal data, dynamic performance of depression behavior is comprehensively captured, and under the adjustment of a multi-modal bridge fusion module, data between modes is fully interacted, the utilization rate of multi-modal data is improved, and therefore the accuracy and efficiency of automatic depression evaluation are improved.

The invention relates to a spatial light field coding device and a spatial light field coding method for light field beam driving guidance, in particular to a spatial light field coding device and a spatial light field coding method based on orthogonal time division multiplexing. The technical problems that an existing space coding technology for light field beam driving guiding is low in coding frequency and poor in reliability are solved. The device comprises a photoelectric sensor, a laser, a collimating lens, a composite prism, a projection lens and a spatial light modulator. The composite prism comprises a lower prism and an upper prism; output light of the laser is collimated by the collimating lens to form collimated laser, the collimated laser is projected to the spatial light modulator through the composite prism to perform light field intensity coding modulation, and the modulated light field returns to the composite prism and is projected by the projection lens to form an imaging spatial light field; and the photoelectric sensor is arranged on the to-be-guided device and used for receiving the code of the image space light field, decoding the code to obtain the position of the to-be-guided device and adjusting the steering engine according to the position to enable the to-be-guided device to move towards a preset position.

Disclosed is a hierarchical spatial resolution codec that adaptively adjusts the representations of immersive audio content as the target bandwidth for delivering the audio content changes. The audio content may be represented by an adaptive number of content types such as channels / objects, higher-order ambisonics (HOA), and encoded by adaptive spatial coding techniques to support the target bitrate of a transmission channel or user. Adaptive spatial coding techniques may include adaptive channel / object spatial encoding techniques to generate an adaptive number of channels / objects, and adaptive HOA spatial encoding or HOA compression techniques to generate an adaptive order of the HOA. The adaptation may be a function of the target bitrate that is associated with a desired quality, and an analysis that determines the priority of the channels, objects, and HOA. High priority channels / objects may be encoded into a high quality bit-stream while low priority channels / objects may be converted and encoded as HOA.

Systems and methods for preserving headphone rendering mode (HRM) in object clustering are described. In an embodiment, an object-based audio data processing system includes a processor configured to receive a plurality of audio objects, wherein an audio object of the plurality of audio objects is associated with respective object metadata that indicates respective spatial position information and an HRM; determine a plurality of cluster positions by applying an extended hybrid distance metric to a spatial coding algorithm to calculate a partial loudness for each of the audio objects; render the audio objects to the cluster positions to form a plurality of clusters by applying the extended hybrid distance metric to the spatial coding algorithm to calculate object-to-cluster gains; and transmit the clusters to a spatial reproduction system.

The invention relates to the technical field of deep learning and meteorological science crossing, in particular to a deep learning rainfall prediction enhancement method fusing a graph structure and frequency domain characteristics and an integration framework, and the method comprises the steps: obtaining rainfall data of a historical time-space sequence as input data; inputting the input data into a spatial coding module to extract initial spatial features; inputting the initial spatial features into a graph frequency fusion enhancement module to generate enhanced features; wherein the graph-frequency fusion enhancement module is used for enhancing the expression capability of a local space structure and multi-scale frequency domain features of a precipitation field by combining Fourier frequency domain feature extraction and spatial topology modeling of a graph neural network; and inputting the enhanced features into a backbone network module of a rainfall prediction model, carrying out spatio-temporal evolution modeling, and outputting a rainfall prediction result of a future spatio-temporal sequence. According to the method, the accuracy and reliability of heavy rainfall prediction can be improved, and more powerful technical support is provided for weather forecast and disaster early warning.

The embodiment of the invention discloses a space-time unified coding method and device, computer equipment and a storage medium. The method comprises the steps of obtaining space information, time information and equipment information corresponding to target equipment; encoding the spatial information to obtain a spatial code corresponding to the target equipment; encoding the time information to obtain a time code corresponding to the target equipment; encoding the equipment information to obtain an equipment code corresponding to the target equipment; generating a space-time coding identifier corresponding to the target equipment according to the space code, the time code and the equipment code; initial data collected by the target device are obtained, target data are generated according to the combination of the space-time coding identifier and the initial data, and the space-time coding identifier is used for indexing the initial data contained in the target data and / or conducting data alignment between the target data and other data containing the preset space-time coding identifier. Therefore, the space-time coding identification can be used for indexing and aligning the data.

The invention provides a land resource information monitoring and management system, and the system comprises a three-dimensional spatio-temporal data collection module which collects overground and earth surface spatio-temporal data, integrates spatio-temporal data of different spatial scales and time scales, and forms a three-dimensional spatio-temporal data set; the data fusion module is used for preprocessing the three-dimensional space-time data set and carrying out multi-source fusion on the three-dimensional space-time data set; the spatial code data association module is used for generating spatial codes according to the multi-source data fused by the data fusion module in combination with geographic coordinates, timestamps, resource types and region identifiers, and associating different service data through the spatial codes to form a data network; the space tent tracking module is used for calculating the total resource amount and the resource variation amount according to the three-dimensional spatio-temporal data, analyzing the resource flow direction according to space codes and dynamically tracking the city-county-rural three-level resource state; and the visualization module is used for carrying out visualization display on data and analysis results of all the modules.

The invention discloses a wiring-free Carplay reversing image wireless transmission method and a wiring-free Carplay reversing image wireless transmission system, and belongs to the technical field of video coding. When a reversing mode is started, a wireless transmission link monitoring thread is started, a wireless transmission link performance index is acquired, and an acceptable target code rate of a current reversing image frame is given; after acquiring each frame of image, the reversing image acquisition equipment analyzes to obtain an initial block-level coding precision matrix, performs inter-frame smoothing and variation amplitude constraint on the coding precision of each spatial position in the matrix, and performs differential coding on each spatial coding block according to a final block-level coding precision matrix to obtain a final block-level coding precision matrix; the method comprises the following steps: generating a coding data stream adaptive to the bearing capacity of a current link, sending the coding data stream to a CarPlay terminal through a wireless transmission link, obtaining a monitoring result in a transmission process, dynamically and finely adjusting the target code rate and block-level precision distribution of a subsequent frame, and reducing the impact of queue backlog and frame-level loss of a sending end on the real-time performance and accuracy of a reversing image from the source.

The application provides a cross-domain flood prediction system based on spatial coding and coupled bidirectional LSTM, relates to the technical field of flood prediction, and constructs a three-layer collaborative architecture of a perception node, an edge gateway and a cloud engine. The perception node collects multi-modal hydrological data and pre-processes; the edge gateway generates a basin grid through spatial interpolation of discrete sensor data, extracts multi-scale spatial features through spatial pyramid pooling and reduces dimension to a spatial coding vector; the cloud engine initializes the vector as a cell state of a forward long short-term memory network, and injects the forward hidden state into each gating unit of a reverse long short-term memory network through a coupling matrix in real time, realizes forward and reverse information collaboration, and finally outputs a future runoff prediction value. The application can retain the spatial heterogeneity structure of the basin, fully utilize future weather forecast information, and significantly improve the cross-region prediction accuracy and system disaster resistance.

The invention belongs to the technical field of data processing, is applied to medical scenes and financial scenes, and discloses a resource scheduling method, device, equipment and medium based on a multi-level thermodynamic map, and the method comprises the following steps: obtaining business data of each level of an administrative region, the business data comprises spatial position information and resource state indexes of each level of the administrative region; converting the spatial position information of each level of the administrative region into geographic coordinate information through a spatial coding engine, and constructing a spatio-temporal data set of each level of the administrative region based on the geographic coordinate information; a map of each level of the administrative region is loaded, hierarchical rendering is performed on the map of each level of the administrative region according to the resource state indexes in the spatio-temporal data set, and a thermodynamic distribution diagram of each level of the administrative region is generated; and the resource state of each level of the administrative region is identified based on the generated thermodynamic distribution diagram, and the corresponding resource scheduling configuration instruction is generated based on the identified resource state, so that the resource scheduling efficiency of the service scene is effectively improved.

The application discloses an automatic driving space planning enhancement method based on semantic region focusing, comprising the following steps: acquiring input information, including images of an automatic driving scene and texts of a space planning problem; designing a dynamic text library and constructing cross-modal correlation mapping of the text library and the images to generate a region-activated attribution graph; performing spatial visual coding on the images and fusing the attribution graph to construct an explicit spatial coding prompt, constructing a geometric feature enhancement prompt and combining the explicit spatial coding prompt to generate a structured prompt; inputting the input information and the structured prompt into a multi-modal large model to obtain an answer to the space planning problem; and introducing region attention prior through the explicit visual prompt and the text prompt, improving the attention degree of an automatic driving model to a key semantic region in a complex automatic driving scene, thereby enhancing the space planning capability of the automatic driving system.