33 results about "Coverage density" patented technology

The invention discloses a geological structure three-dimensional reconstruction and analysis system based on multi-modal data fusion, and relates to the technical field of geological exploration information. The method comprises the following steps: acquiring and normalizing multi-source heterogeneous geological observation data through a preprocessing module, and generating basic data and a data coverage density map; the adaptive fusion module performs analysis to generate a dynamic adaptive weight field, and performs weighted fusion on the data to obtain a primary three-dimensional geological data volume; the construction and optimization module constructs a three-dimensional space residual field in a vector form based on the difference between the fusion body and the original data, and deduces the geometric morphology of the geological interface; the reconstruction module ensures the consistency between model layers through hierarchical bidirectional belief propagation processing; the quantification module calculates an uncertainty index; and the visualization module finally generates a three-dimensional geological scene. According to the method, the problems of conflict resolution and reliability evaluation in multi-source geological data fusion are effectively solved, and the precision of the three-dimensional geological model and the engineering decision support capability are improved.

The invention relates to the technical field of parameter monitoring, in particular to a body surface parameter monitoring method and system for meat duck breeding, and the method comprises the steps: collecting the original image data of a meat duck epidermis region, and synchronously obtaining an environment illumination intensity parameter; performing dynamic illumination compensation processing on the original image data to generate a standard illumination epidermis image; performing feather-epidermis region segmentation on the standard illumination epidermis image to generate a feather distribution mask and an epidermis exposure region coordinate set; extracting multi-dimensional physical sign parameters based on the epidermis exposure area coordinate set, wherein the multi-dimensional physical sign parameters comprise feather gap epidermis temperature, a feather coverage density index and an epidermis texture roughness coefficient; inputting the multi-dimensional physical sign parameters into a pre-trained growth state evaluation model, and outputting a physical sign anomaly probability value and a corresponding anomaly type code; according to the invention, feather dynamic shielding and ambient light interference are effectively overcome, a real-time linkage mechanism of physical sign parameters and feeding strategies is established, and the intelligent level and economic benefits of meat duck breeding are significantly improved.

The invention provides a scene reconstruction and automatic correction method using an unmanned aerial vehicle, and belongs to the technical field of scene reconstruction and correction, and the method comprises the steps: carrying out the global data collection of different sub-regions, synchronously recording the flight path information of each round, and carrying out the timestamp alignment processing of sensor data; constructing a dynamic error compensation model, and correcting motion blur, environmental interference and trajectory deviation errors in the collected data by combining a dynamic parameter set of random flight of each round of the unmanned aerial vehicle; integrating multiple rounds of data after correction, establishing spatial mapping of subarea textures and point clouds, and generating an initial three-dimensional model; and constructing a closed-loop correction mechanism based on the GNSS absolute coordinate of the unmanned aerial vehicle and the ground high-precision control point, calculating the deviation of the initial three-dimensional model, and if the deviation exceeds a threshold value, reversely adjusting the random flight coverage density, the flight round and the error compensation coefficient until the deviation reaches the standard. And the precision of scene reconstruction and correction is improved.

The invention relates to the technical field of spartina alterniflora image recognition, in particular to a spartina alterniflora distribution condition monitoring and early warning method based on image recognition, which comprises the following steps: extracting vegetation features by utilizing Sentinel-1 radar data and Sentinel-2 optical remote sensing data, and realizing spartina alterniflora region classification in combination with a TWDTW model to obtain a first spartina alterniflora region; calculating an ecological factor and a water conservancy factor of the first spartina alterniflora region, and generating a graded region through weighted summation; and regularly using the unmanned aerial vehicle to shoot ground images of a first spartina alterniflora area according to the priority of the graded areas, and using the target detection model to identify distribution change of the spartina alterniflora, generating indexes of a coverage area change rate, a diffusion rate and a coverage density change rate, and forming a dynamic early warning report. According to the method, the spartina alterniflora area is identified and the diffusion trend of the spartina alterniflora is analyzed in real time, so that the monitoring accuracy of the distribution condition of the spartina alterniflora is improved, and a scientific decision basis is provided for ecological environment protection and spartina alterniflora management.

The invention provides a wetland ecological restoration effect comprehensive evaluation method and system, and belongs to the technical field of big data analysis. According to the method, real-time data of multi-dimensional ecological elements such as water turbidity, vegetation coverage density and soil acidity and alkalinity are collected, a multi-dimensional element correlation matrix based on time sequence correlation is constructed, and the dynamic change frequency of each element is captured. Core influence factors are screened by quantifying the contribution degree of each element to ecological restoration, weight distribution corresponding to clustered element clusters is determined, and then an accurate quantitative score is generated and mapped into a clear restoration condition level through multiplication and summation of inter-cluster correlation coefficients and weights item by item. Meanwhile, an evolution track is compared and analyzed by combining historical time sequence data, and the weight is dynamically adjusted to ensure that evaluation adapts to an actual repair process. The comprehensive, dynamic and accurate evaluation of the wetland ecological restoration effect is realized, scientific support is provided for the accurate adjustment of the restoration strategy, and the fine management of the wetland and the sustainable development of the natural environment are guaranteed.

The present disclosure provides a forest terrain remote sensing mapping method and system for smart forestry, which relates to the field of data processing. The method includes: obtaining a remote sensing image of the forest area to be measured; obtaining the elevation value of each monitoring station in the forest area to be measured and the grayscale value at the corresponding position in the remote sensing image, and establishing an elevation-grayscale relationship model; for each partition of the forest area to be measured, determining the vegetation coverage density of the partition based on the grayscale value of each pixel point in the remote sensing image corresponding to the partition; determining the theoretical elevation value corresponding to the grayscale value of each pixel point in the remote sensing image corresponding to the partition based on the elevation-grayscale relationship model, and correcting the measured elevation value based on the vegetation coverage density and the theoretical elevation value to obtain a weighted elevation value; determining the seasonal weight based on the temperature of the forest area to be measured, and determining the target elevation value of the partition based on the seasonal weight and the weighted elevation value. The present disclosure can improve the accuracy of forest elevation data.

The invention relates to the technical field of photovoltaic panel control, and discloses a photovoltaic panel output control method and system, and the method comprises the steps: monitoring the accumulated snow thickness and dust coverage density of the surface of a photovoltaic panel in real time, and calculating the actual power generation amount and power generation efficiency loss; dynamically adjusting the inclination angle of the photovoltaic panel or triggering self-cleaning operation; and calculating the optimal inclination angle of the photovoltaic panel by using an improved particle swarm optimization algorithm. In the prior art, when a photovoltaic panel is covered by accumulated snow or dust, the coverage state cannot be monitored in real time and the inclination angle cannot be dynamically adjusted, and especially under complex weather conditions, the power generation amount is greatly reduced due to low efficiency or too high cleaning energy consumption of a system. Through real-time coverage state monitoring and inclination angle dynamic adjustment based on particle swarm optimization, efficient output control of the photovoltaic panel inclination angle control motor is realized, so that the problems that the power generation efficiency is influenced by coverings and the cleaning energy consumption is too large are avoided, and the energy utilization rate of the system is improved.

An image collection method and an image capturing device. The method comprises: obtaining first coverage density data of one or more first images based on captured first images of a target object to determine one or more re-shooting positions; rendering one or more shooting position suggestion marks corresponding to the one or more re-shooting positions in an image corresponding to a real space displayed in real-time by the image capturing device; and when a current position and a current viewing angle of the image capturing device match a target shooting position suggestion mark, automatically capturing a target second image corresponding to the target shooting position suggestion mark toward the target object. Thereby, the present disclosure provides a method that can effectively guide users to complete image collection required for 3D object modeling.

The present invention discloses a method and system for analyzing surveying and mapping information data for land planning and surveying projects, specifically relating to the technical field of topographic surveying and mapping, and is used to resolve the technical defect of missing surface point clouds caused by fixed laser parameters in vegetation-covered areas. The method obtains near-infrared spectral images, polarization characteristics, and aircraft pitch angles of vegetation areas in real time; calculates vegetation coverage density and inverts leaf inclination distribution entropy; couples multiple scattering effects with geometric propagation characteristics to generate actual penetration path length; dynamically increases pulse energy based on a coverage density threshold; increases emission frequency in stages according to the proportion of penetration path excess; and uses optimized parameters to emit laser beams to generate a high-integrity surface point cloud dataset. The method solves the terrain blind spot problem caused by vegetation occlusion and provides terrain model support for land planning.

This invention provides a dust pollution detection method and system based on image recognition, belonging to the field of image processing technology. The method includes: acquiring a group of dust images of a region to be detected, and determining a target image from the dust image group based on a preset filtering rule; performing preprocessing on the target image to obtain a binarized dust image; determining an improved grayscale mean to characterize the global coverage density of the dust based on the binarized dust image, and determining a fractal dimension to characterize the local morphological complexity of the dust using an improved box counting method based on the binarized dust image; constructing a pollution factor based on the improved grayscale mean and the fractal dimension, and assessing the dust pollution level based on the pollution factor. This invention achieves a rapid, stable, and structurally discriminative quantitative assessment of dust pollution status in semi-enclosed environments such as tunnels.

The invention discloses a method and system, and the method comprises the following steps: obtaining the thickness of a corner protector according to the sharpness of a single sharp node, and obtaining the volume of the corner protector according to the shape of a region covered by the corner protector and the thickness of the corner protector; when the total volume of the required angle bead material is within a threshold value, comparing angle bead covering density values obtained before and after angle bead covering and obtaining a corresponding collapse coefficient, and obtaining the surface area of the plastic package material according to the product of the collapse coefficient and a correction coefficient of a proportion value of each area of the edge occupying the total area; otherwise, determining different thicknesses of the corresponding nodes on the packaging material according to the score difference degree; a packaging part combination is obtained under the rule that the shape and the size of a packaging part are not changed and edge nodes of overlapped packaging parts are within a threshold value, a cuboid is constructed according to the tangent line of the outermost side, and the packaging group with the length, width and height parameters within the threshold value and the minimum gap total size is selected to conduct secondary packaging on the packaging part. The method has the characteristic of improving the accuracy and efficiency of the packaging service.

The invention discloses a forest land remote sensing extraction deep learning method fusing spatial and semantic information, and the method comprises the steps: constructing an S2-IFNet model based on a deep learning model, firstly constructing a dual-branch feature extraction module to extract the spatial and semantic features of a forest land, and then adding boundary features to enhance the S2-IFNet model so as to enhance the edge features of the forest land; then an attention mechanism module is introduced to further improve the forest land feature capturing capability of the S2-IFNet model, and finally a boundary feature enhanced space-semantic information fusion strategy is introduced to effectively fuse the space and semantic features of the forest land so as to improve the extraction capability of the forest land and improve the extraction efficiency of the forest land. The problems of adhesion of extraction results and serious extraction leakage phenomena caused by different coverage density degrees and more confused ground features of the forest land are effectively solved. More importantly, the method provided by the invention not only can better extract the forest land range, but also is suitable for extracting different types of forest lands in different regions, and has better robustness.

This invention belongs to the field of shear wave exploration technology and discloses a design method for a shear wave observation system. The method involves: acquiring target layer parameter information in the exploration area; determining the shear wave elements of the shear wave observation system based on the target layer parameter information; determining the shear wave receiver spacing of the shear wave observation system based on the target layer parameter information; quantitatively calculating the maximum shear wave offset based on the time-distance curves of the direct shear wave, refracted shear wave, and reflected shear wave; verifying the effectiveness of the shear wave elements in step S2, the shear wave receiver spacing in step S3, and the maximum shear wave offset in step S4 based on forward modeling of the shear wave equation and actual shear wave seismic records acquired through field experiments; and calculating the coverage number and coverage density of the shear wave observation system based on the signal-to-noise ratio of the actual shear wave data and the effective offset range of the shear wave. This invention obtains high-quality shear wave seismic data through qualitative and quantitative design of a nine-component three-dimensional shear wave observation system. This invention is applicable to the design of shear wave observation systems.

One variation of a method includes: detecting a first location of a user carrying a mobile pack proximal a first scan zone; accessing a first sequence of images representing trees in the first scan zone; characterizing a first coverage density of the first scan zone based on visual features detected in the first sequence of images; in response to the first coverage density exceeding a threshold coverage density, alerting the user of scan completion; detecting a second location of the user proximal a second scan zone; accessing a second sequence of images representing trees in the second scan zone; characterizing a second coverage density of the second scan zone based on visual features detected in the second sequence of images; and, in response to the second coverage density falling below the threshold coverage density, prompting the user to increase a traversal speed for a third scan zone.

The invention relates to the technical field of thickness measurement, in particular to an X-ray thickness gauge material compensation method, which comprises the following steps of: setting a plurality of acquisition angles of an X-ray source and a detector in a range of-15 degrees to + 15 degrees, performing X-ray transmission on a measured object at each angle, and recording a transmission intensity value received by the detector. The X-ray source and the detector are set to carry out multi-angle acquisition in the angle range of-15 degrees to + 15 degrees, the coverage density of information of a detected object in different directions is improved, the analysis capability of a complex structure cross section is enhanced, and on the basis of multi-angle attenuation projection data, the detection accuracy is improved. According to the method, the cross section area of the measured object is discretized into the two-dimensional pixel grid, so that consistent modeling of the heterogeneous object is realized, initial modeling deviation caused by prior insufficiency of materials is avoided, and subsequent iterative correction operation is more accurate through geometric mapping of a pixel path and a projection path.

The present application relates to the technical field of underground pipe network monitoring and early warning, and particularly relates to an underground pipe network partition visualization and safety early warning method and system. The method comprises: obtaining a map service source, an underground pipeline service, GIS data, a BIM model, equipment data, historical monitoring data, historical maintenance records, inspection logs and accident reports, performing data alignment and data standardization processing to obtain pipe network topology structure information; performing key node, valve, connection relationship and monitoring coverage density correlation processing to obtain partition objects; performing connection processing to obtain multi-source heterogeneous monitoring data, and performing time sequence feature coding, spatial feature coding, semantic text data feature coding and cross-modal fusion processing to obtain fusion feature vectors; and then performing confidence interval, intelligent alarm instruction, mutation alarm instruction, maximum clustering distance and connection relationship joint processing to obtain correlation alarm state information. The present application can effectively improve the accuracy and timeliness of safety early warning.

The invention provides an earthquake visual early warning method combining monocular depth estimation and an optical flow method. According to the method, through fusion of monocular depth estimation and a robust optical flow algorithm suitable for a traffic scene, inversion of ground acceleration at a street lamp type equipment base is realized, a distributed earthquake monitoring network can be constructed based on inversion data, the industrial pain points of high deployment cost and insufficient coverage density of a traditional seismograph are solved, and the method is suitable for popularization and application. And a solution with high cost performance is provided for urban earthquake early warning.

The invention relates to an RGB terrain loading method and device for Cesium and electronic equipment. The method comprises the steps of converting RGB tile data into height data, generating a geometric error graph based on the height data, screening out control points from the geometric error graph, generating a terrain triangulation network based on the control points, and performing terrain drawing according to the terrain triangulation network. The control points are screened out from the generated geometric error graph, and the more accurate terrain triangulation network is generated based on the screened control points, so that the defect of data distortion in the prior art is overcome, in addition, the coverage density is improved by increasing the number of the terrain triangulation networks, the inclination angle between the terrain triangulation networks is reduced, and the accuracy of the terrain triangulation network is improved. Therefore, the problems of sharpness and unsmooth transition of the existing triangulation network are solved.

The invention provides a nanometer quick-freezing fresh-locking method based on a superposed space electric field, and relates to the technical field of freezing fresh-locking, and the method comprises the following steps: obtaining a three-dimensional space arrangement of muscle fibers in a nanometer quick-freezing liquid, and identifying a fiber arrangement included angle and a fiber coverage density of the muscle fibers, collecting the initial electric field direction of the superposition space electric field and the polarization intensity of each area to obtain a polarization intensity mean value; determining a one-way electric field coverage rate according to whether the fiber arrangement included angle exceeds a preset one-way electric field coverage range threshold value or not, and determining polarization intensities of fibers with different orientations according to the one-way electric field coverage rate and the fiber coverage density; and an electric field switching period is generated based on the adjusted polarization intensity distribution, the difference value change trend of the energy storage value in the fiber and the energy storage value in the fiber gap after the electric field switching period is executed is analyzed, the balance degree of energy storage inside and outside the fiber is evaluated, and whether cold energy storage exceeds a preset range or not is monitored.

This invention relates to an evaluation method, apparatus, electronic device, and storage medium for a seismic observation system. The method involves acquiring a geological model of an exploration area; determining the maximum shot-receiver distance for each subsurface layer and, based on the maximum shot-receiver distance, determining the seismic observation system corresponding to each subsurface layer; identifying any subsurface layer as a target layer; obtaining the imaging area projected onto the surface of a target point on the target layer using a reverse ray tracing method, and determining the area of ​​the imaging area; determining the coverage frequency of the imaging area based on the seismic observation system corresponding to the target layer; and determining the coverage density of the imaging area based on the coverage frequency and the area of ​​the imaging area. This coverage density is used to evaluate the seismic observation system. Specifically, the method involves targeted statistical analysis of coverage density based on the maximum offset corresponding to different subsurface layers, using coverage density to determine the rationality of the observation system, thereby improving the evaluation rationality.

This invention belongs to the field of agricultural image processing technology and discloses a sensor distribution optimization method, a plant stem diameter measurement method, and a device for optical tomography of plant stems. The optimization method mainly involves determining the optimal detection range, and then measuring the plant stem diameter within this range. The specific steps are: 1. Acquiring the background light signal without placing the plant; 2. Acquiring the light signal from the plant stem; 3. Constructing a projection data matrix, obtaining the column vector of the projection data matrix through the background light signal and the plant stem light signal; 4. Performing image reconstruction to obtain a cross-sectional image of the plant stem; 5. Filtering and thresholding the cross-sectional image, calculating the total gray area, and then calculating the equivalent diameter of the plant stem using the total gray area. This method not only improves the light coverage density and distribution uniformity of the measurement area but also enhances measurement accuracy.

The invention relates to a throwing device capable of regulating and controlling explosion energy distribution, and belongs to the technical field of ammunition. The device comprises a cabin opening mechanism, a shell, a base and a blasting unit, wherein the cabin opening mechanism is used for radially separating the shell to expose the blasting unit; the blasting unit comprises a blasting assembly, the blasting assembly comprises a throwing mechanism and a throwing body, the throwing body is installed on the throwing mechanism, the throwing mechanism is used for throwing the throwing body, and the throwing body is used for covering a target area. The cabin opening mechanism and the scattering mechanism are arranged to be used for cabin opening and scattering of the scattering body correspondingly, after cabin opening and scattering of the scattering body are achieved, the cabin opening mechanism and the scattering mechanism do not disintegrate and can serve as a whole with the base to continue to cover the target area, especially the center area of the target area can be covered, and therefore the target area can be more effectively covered. Therefore, the coverage density of the target area is improved, and the coverage effect is improved.

This invention provides a method for scene reconstruction and automatic correction using unmanned aerial vehicles (UAVs), belonging to the field of scene reconstruction and correction technology. It involves comprehensive data acquisition across different sub-regions, synchronously recording flight trajectory information for each round, and aligning sensor data with timestamps. A dynamic error compensation model is constructed, combining the dynamic parameter set of each round of random flight by the UAV to correct motion blur, environmental interference, and trajectory deviation errors in the acquired data. The corrected multi-round data is integrated to establish a spatial mapping between sub-region textures and point clouds, generating an initial 3D model. A closed-loop correction mechanism is constructed based on the UAV's GNSS absolute coordinates and high-precision ground control points to calculate the deviation of the initial 3D model. If the deviation exceeds a threshold, the random flight coverage density, flight rounds, and error compensation coefficients are adjusted in reverse until the deviation meets the standard. This improves the accuracy of scene reconstruction and correction.

The invention provides a forest terrain remote sensing surveying and mapping method and system for smart forestry, and relates to the field of data processing. The method comprises the following steps: acquiring a remote sensing image of a forest region to be measured; the elevation value of each monitoring station in the forest region to be measured and the gray value of the monitoring station at the corresponding position of the remote sensing image are acquired, and an elevation-gray relation model is established; for each partition of the forest region to be measured, determining the vegetation coverage density of the partition based on the gray value of each pixel point in the remote sensing image corresponding to the partition; determining a theoretical elevation value corresponding to the gray value of each pixel point in the remote sensing image corresponding to the partition based on an elevation-gray relation model, and correcting the measured elevation value based on the vegetation coverage density and the theoretical elevation value to obtain a weighted elevation value; and determining a seasonal weight based on the air temperature of the forest region to be measured, and determining a target elevation value of the partition based on the seasonal weight and the weighted elevation value. According to the invention, the accuracy of the elevation data of the forest region can be improved.

The invention discloses a monitoring point layout method, device and equipment, a medium and a product. The method comprises the following steps: acquiring a monitoring point data set and an event data set corresponding to a target area; determining a monitoring thermal grid map corresponding to the target area according to the monitoring point data set; the monitoring heating power is used for representing the coverage density of monitoring points in the grid units; determining an event thermal grid map corresponding to the target area according to the event data set; the event heat is used for representing the risk level intensity of the event occurring in the grid unit; performing intersection negation operation on the monitoring thermal grid map and the event thermal grid map to obtain a to-be-monitored area corresponding to the target area; and determining a monitoring point layout scheme corresponding to the target area according to the to-be-deployed and controlled area. Through the technical scheme of the invention, the monitoring point location layout can be optimized based on the thermal grid map, and the event risk area is reasonably covered while the investment of redundant equipment is reduced.

The invention provides a design method, system, equipment and medium based on a residual gas well underground combined mining three-dimensional observation system, and belongs to the technical field of petroleum geophysical exploration. According to the method, aiming at the problems of complicated geological conditions and insufficient resolution of a conventional observation system in residual gas exploration, observation points are arranged by comprehensively analyzing geological condition factors and technical parameters of the observation system and adopting a multi-dimensional optimization strategy. The method specifically comprises the following steps: constructing a fitness function based on wave field energy coverage maximization, reflection interface coverage and geological parameter mean square error minimization, and carrying out adaptive observation point position design; according to the dimension difference of well-ground combined mining, ground excitation points are finely adjusted to make up a well periphery covering density gap; and on the basis of a static correction principle, according to topographic relief and reservoir heterogeneity characteristics, excitation and receiving parameters are adjusted in a partitioned manner. The method can effectively improve the signal-to-noise ratio and imaging precision of seismic data under complex geological conditions, accurately depicts the residual gas distribution, and improves the exploration efficiency.

The invention discloses a method and system for constructing a space-air-ground three-dimensional cooperative monitoring network system, and relates to the technical field of environment monitoring, and the method specifically comprises the following steps: dividing a monitoring grid aggregation task coordinate calculation density interpolation assignment generation block, correcting a coverage radius to normalize a node attribute generation capability list, and generating a monitoring grid aggregation task coordinate calculation density interpolation assignment generation capability list; screening efficient nodes to construct cost matrix bilateral matching, extracting feature vectors, calculating distances, identifying redundant nodes, removing redundancy, and supplementing candidate output space-ground-air collaborative configuration tables. According to the method, the monitoring block is dynamically generated based on the geographic grid, the coverage radius is corrected in combination with the node direction and the hot area distance, the normalization ability list is constructed, the task grid and the node are matched through the Hungary algorithm, the cost matrix is constructed to recognize the redundant node, self-adaptive matching of the node block is achieved, the resource allocation efficiency is improved, and parameter dependence is reduced; the monitoring blind area is reduced, the coverage density and the energy consumption level are balanced, and the emergency response sensitivity is enhanced.

The invention discloses a surveying and mapping information data analysis method and system for a land planning investigation and survey project, particularly relates to the technical field of topographic surveying and mapping, and is used for solving the technical defect of surface point cloud missing caused by fixed laser parameters in a vegetation coverage area. The method comprises the following steps: acquiring a near-infrared spectrum image, polarization characteristics and an aircraft pitch angle of a vegetation area in real time; calculating a vegetation coverage density and inverting a leaf dip angle distribution entropy value; coupling multiple scattering effects and geometric propagation characteristics to generate an actual penetration path length; dynamically increasing pulse energy based on a coverage density threshold; improving the transmitting frequency in a grading manner according to the penetration path excess proportion; transmitting a laser beam by adopting optimized parameters to generate a high-integrity earth surface point cloud data set; the problem of terrain blind areas caused by vegetation shielding is solved, and terrain model support is provided for land planning.

The application provides a seismic visual early warning method combining monocular depth estimation and an optical flow method. The method realizes inversion of ground acceleration at a base of a street lamp type device by fusing monocular depth estimation and a robust optical flow algorithm suitable for a traffic scene, and can construct a distributed seismic monitoring network based on inversion data, solves the industry pain points of high deployment cost and insufficient coverage density of a traditional seismograph, and provides a high cost-effective solution for city-level earthquake early warning.

The embodiment of the invention provides a dust pollution detection method and system based on image recognition, and belongs to the technical field of image processing. The method comprises the following steps: acquiring a dust image group of a to-be-detected area, and determining a target image from the dust image group based on a preset screening rule; performing preprocessing on the target image to obtain a dust binary image; determining an improved gray average value used for representing the global coverage density of the dust based on the dust binary image, and determining a fractal dimension used for representing the local form complexity of the dust by adopting an improved box counting method based on the dust binary image; and constructing a pollution factor based on the improved gray average and the fractal dimension, and evaluating the dust pollution level according to the pollution factor. According to the scheme provided by the invention, the dust pollution state of the tunnel type semi-closed environment can be rapidly and stably quantitatively evaluated with structural discrimination.