Bloodstain form automatic analysis and dripping parameter calculation system based on machine vision

Abstract

The invention relates to the field of bloodstain detection. By providing a bloodstain form automatic analysis and dripping parameter calculation system based on machine vision, the method comprises the following steps: performing multi-modal image acquisition on a target bloodstain to obtain a high-resolution three-dimensional image comprising a microscopic crystal structure; dynamic texture feature extraction is carried out, and dynamic texture features associated with the blood drying time are generated; performing multi-modal fusion processing in combination with the environmental parameters and the dynamic texture features, constructing a time-texture mapping library, and generating a probability distribution model of bloodstain formation time; performing formation time calculation according to the probability distribution model, generating the formation time of the target bloodstain, and correcting the parameter calculation result of the dripping height and angle based on the formation time to obtain corrected dripping height and angle parameters; the three-dimensional bleeding point position is reconstructed through a reverse projection algorithm, so that the problem of limitation of a related bloodstain analysis technology on time estimation and parameter correction is solved, and the accuracy and scene adaptability of bloodstain analysis are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of bloodstain detection, and particularly to an automatic analysis system for bloodstain morphology and a calculation system for dripping parameters based on machine vision. Background Art

[0002] With the continuous development of artificial intelligence and computer vision technologies, bloodstain morphology analysis plays an increasingly important role in forensic identification, crime scene reconstruction, and other fields. Accurate estimation of bloodstain formation time and calculation of dripping parameters are one of the key steps in restoring the crime process and also technical difficulties in forensic physical evidence analysis.

[0003] However, traditional bloodstain analysis technologies have significant limitations in time estimation and parameter correction. Related technologies rely on empirical judgment and two-dimensional morphology measurement, ignoring the dynamic changes in the microscopic crystal structure during the blood drying process, resulting in large time prediction errors; at the same time, the influence of environmental temperature and humidity on blood viscosity and diffusion speed is not fully considered, making it difficult to accurately correct the dripping height and contact angle parameters; in addition, the three-dimensional bleeding point reconstruction technology based on a static physical model is prone to spatial positioning errors due to differences in bloodstain drying degree because of the lack of a timeliness compensation mechanism. Summary of the Invention

[0004] Based on this, it is necessary to provide an automatic analysis system for bloodstain morphology and a calculation system for dripping parameters based on machine vision to address the limitations of related bloodstain analysis technologies in time estimation and parameter correction, thereby improving the accuracy and scene adaptability of bloodstain analysis.

[0005] In a first aspect, the present application provides an automatic analysis system for bloodstain morphology and a calculation system for dripping parameters based on machine vision, which includes:

[0006] A multi-modal image acquisition module for performing multi-modal image acquisition on the target bloodstain to obtain a high-resolution three-dimensional image including the microscopic crystal structure;

[0007] A dynamic texture feature extraction module for extracting dynamic texture features from the high-resolution three-dimensional image based on a transfer learning model to generate dynamic texture features associated with the blood drying time;

[0008] A time-texture modeling module for performing multi-modal fusion processing by combining environmental parameters and dynamic texture features to construct a time-texture mapping library and generate a probability distribution model of the bloodstain formation time;

[0009] A time estimation module for calculating the formation time according to the probability distribution model to generate the formation time of the target bloodstain, and correcting the calculation results of the dripping height and angle parameters based on the formation time to obtain the corrected dripping height and angle parameters;

[0010] A three-dimensional reconstruction module, which is used to reconstruct the three-dimensional bleeding point position through the inverse projection algorithm based on the corrected dripping height and angle parameters.

[0011] Furthermore, the formation time is calculated according to the probability distribution model to generate the formation time of the target bloodstain, and the calculation results of the dripping height and angle parameters are corrected based on the formation time to obtain the corrected dripping height and angle parameters, including:

[0012] Extract the dry time probability distribution data of the target bloodstain from the probability distribution model to generate a time estimate value and confidence interval parameters;

[0013] Based on the high-resolution three-dimensional image, the initial dripping height parameter and the initial contact angle parameter are calculated through the hydrodynamic model;

[0014] According to the time estimate value, perform time decay compensation calculation on the initial dripping height parameter and the initial contact angle parameter to generate the corrected dripping height parameter and contact angle parameter;

[0015] Based on the confidence interval parameters, verify the error boundary of the corrected dripping height parameter and contact angle parameter through Monte Carlo simulation to generate the corrected dripping height and angle parameters.

[0016] Furthermore, according to the time estimate value, perform time decay compensation calculation on the initial dripping height parameter and the initial contact angle parameter to generate the corrected dripping height parameter and contact angle parameter, including:

[0017] Based on the variation law of blood viscosity with dry time, establish a mapping relationship between time and viscosity decay coefficient;

[0018] According to the time estimate value, match the corresponding viscosity decay coefficient from the mapping relationship;

[0019] Based on the viscosity decay coefficient, perform time-varying viscosity compensation calculation on the initial dripping height parameter to generate the corrected dripping height parameter;

[0020] Based on the viscosity decay coefficient, perform time-varying viscosity compensation calculation on the initial contact angle parameter to generate the corrected contact angle parameter;

[0021] Perform hydrodynamic simulation verification on the corrected dripping height parameter and contact angle parameter to ensure that they meet the physical constraint conditions of the bloodstain diffusion pattern.

[0022] Furthermore, based on the confidence interval parameters, verify the error boundary of the corrected dripping height parameter and contact angle parameter through Monte Carlo simulation to generate the corrected dripping height and angle parameters, including:

[0023] Generate a random parameter combination of the dripping height and the contact angle based on the confidence interval parameters;

[0024] Perform a hydrodynamic simulation on the random parameter combination to generate a simulation result of the blood stain diffusion pattern;

[0025] Calculate the parameter error distribution based on the comparison between the simulation result of the blood stain diffusion pattern and the actual morphological data of the target blood stain;

[0026] Determine the error boundaries of the corrected dripping height parameter and the contact angle parameter according to the parameter error distribution, and generate the corrected dripping height and angle parameters based on the error boundaries.

[0027] Further, based on the corrected dripping height and angle parameters, reconstruct the three-dimensional bleeding point position through the back-projection algorithm, including:

[0028] Establish a three-dimensional space coordinate system based on the spatial information of the scene where the blood stain is located;

[0029] Input the corrected dripping height parameter and the contact angle parameter into the back-projection algorithm to generate the initial projection parameters of the blood stain movement trajectory;

[0030] Perform a back-projection calculation on the blood stain movement trajectory based on the initial projection parameters and the hydrodynamic model to generate a three-dimensional space trajectory path;

[0031] Determine the three-dimensional coordinates of the bleeding point according to the intersection point of the three-dimensional space trajectory path and the scene physical constraint conditions;

[0032] Verify whether the three-dimensional coordinates conform to the hydrodynamic law of the blood stain diffusion pattern, and if not, iteratively adjust the projection parameters.

[0033] Further, determine the three-dimensional coordinates of the bleeding point according to the intersection point of the three-dimensional space trajectory path and the scene physical constraint conditions, including:

[0034] Perform a three-dimensional modeling on the physical boundary conditions of the scene where the blood stain is located to generate a scene constraint model;

[0035] Perform a geometric collision detection based on the three-dimensional space trajectory path and the scene constraint model to generate a set of candidate intersection points;

[0036] Verify the hydrodynamic law for the set of candidate intersection points, and screen out the intersection points that meet the blood stain diffusion speed and direction constraints;

[0037] Based on the screened intersection points, determine the optimal three-dimensional coordinates of the bleeding point through a spatial clustering algorithm.

[0038] Further, perform a multi-modal fusion process by combining environmental parameters and dynamic texture features, construct a time-texture mapping library, and generate a probability distribution model of the blood stain formation time, including:

[0039] Collect temperature and humidity parameters of the environment where the target bloodstain is located to generate environmental parameter data;

[0040] Perform feature vector splicing on the dynamic texture feature and environmental parameter data to generate a fused feature vector;

[0041] Based on the drying time labels of historical bloodstain samples, historical dynamic texture features, and historical environmental parameter data, construct a time-texture mapping library;

[0042] Based on the time-texture mapping library, generate a probability distribution model of the bloodstain formation time through Bayesian regression model training.

[0043] Furthermore, based on the drying time labels of historical bloodstain samples, historical dynamic texture features, and historical environmental parameter data, construct a time-texture mapping library, including:

[0044] Label the drying time labels of historical bloodstain samples, and synchronously collect the corresponding historical dynamic texture features and historical environmental parameter data;

[0045] Perform noise reduction processing on the historical dynamic texture features to generate standardized dynamic texture features;

[0046] Perform normalization processing on the historical environmental parameter data and the standardized dynamic texture features to generate a standardized time-texture data set;

[0047] Associate and store the standardized time-texture data set with the drying time labels to construct a time-texture mapping library.

[0048] In a third aspect, the present application also provides a computer device, including a memory and a processor. The memory stores a computer program, and is characterized in that when the processor executes the computer program, the steps of any method in the first aspect of the present application are implemented.

[0049] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored, and is characterized in that when the computer program is executed by the processor, the steps of any method in the first aspect of the present application are implemented.

[0050] The technical solution provided by this application includes the following technical effects: By providing an automatic bloodstain pattern analysis and dripping parameter calculation system based on machine vision, the system includes: a multi-modal image acquisition module for performing multi-modal image acquisition on the target bloodstain to obtain a high-resolution three-dimensional image including a microscopic crystal structure; a dynamic texture feature extraction module for extracting dynamic texture features from the high-resolution three-dimensional image based on a transfer learning model to generate dynamic texture features associated with the blood drying time; a time-texture modeling module for performing multi-modal fusion processing by combining environmental parameters and dynamic texture features to construct a time-texture mapping library and generate a probability distribution model of the bloodstain formation time; a time estimation module for calculating the formation time according to the probability distribution model to generate the formation time of the target bloodstain, and correcting the calculation results of the dripping height and angle parameters based on the formation time to obtain the corrected dripping height and angle parameters; a three-dimensional reconstruction module for reconstructing the three-dimensional bleeding point position based on the corrected dripping height and angle parameters through an inverse projection algorithm to solve the limitations of related bloodstain analysis techniques in time estimation and parameter correction, thereby improving the accuracy and scene adaptability of bloodstain analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0052] Figure 1 It is a structural diagram of an automatic bloodstain pattern analysis and dripping parameter calculation system based on machine vision in an embodiment of the present invention;

[0053] Figure 2 It is a flowchart for calculating the formation time according to the probability distribution model to generate the formation time of the target bloodstain, and correcting the calculation results of the dripping height and angle parameters based on the formation time to obtain the corrected dripping height and angle parameters in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0054] In order to make the above objects, features, and advantages of the present application more obvious and understandable, the following will provide a detailed description of the specific implementation manners of the present application with reference to the drawings. Many specific details are set forth in the following description in order to fully understand the present application. However, the present application can be implemented in many other ways different from those described herein. Those skilled in the art can make similar improvements without departing from the connotation of the application. Therefore, the present application is not limited by the specific embodiments disclosed below.

[0055] AsFigure 1 As shown in Figure 1 , the present application provides an automatic bloodstain pattern analysis and dripping parameter calculation system 100 based on machine vision. The system includes:

[0056] A multimodal image acquisition module 101, configured to perform multimodal image acquisition on a target bloodstain to obtain a high-resolution three-dimensional image including a microscopic crystal structure.

[0057] Specifically, select a suitable imaging device, such as an optical microscope, a fluorescence imaging device, a laser scanning confocal microscope, etc., to capture the microscopic structure and surface features of the bloodstain. At the same time, combine structured light scanning or stereovision technology to obtain the three-dimensional shape information of the bloodstain. Collect the surface texture and color information of the bloodstain through a high-resolution optical microscope. Use fluorescence labeling technology to capture the fluorescence signal of specific components in the bloodstain to enhance the visualization of the microscopic crystal structure. Obtain the depth information of the bloodstain through structured light scanning or laser scanning technology and reconstruct its three-dimensional shape. Register the images acquired in different modalities to ensure that the images in each modality are spatially aligned. Then, through an image fusion algorithm, integrate the multimodal data into a unified high-resolution three-dimensional image. Process the acquired images, such as noise reduction and contrast enhancement, to highlight the microscopic crystal structure and surface features of the bloodstain and provide clear image data for subsequent analysis. Use a three-dimensional reconstruction algorithm (such as a voxel grid method or a point cloud reconstruction method) to convert the two-dimensional image data into a three-dimensional model and display the microscopic structure and overall shape of the bloodstain through a visualization tool. Store the acquired high-resolution three-dimensional image data in a database and label information such as the type, location, and environmental parameters of the bloodstain to provide basic data support for subsequent analysis and modeling.

[0058] Through the above steps, it is possible to obtain a high-resolution three-dimensional image of the bloodstain more comprehensively and accurately, and provide high-quality input data for subsequent dynamic texture feature extraction and time-texture modeling.

[0059] A dynamic texture feature extraction module 102, configured to perform dynamic texture feature extraction on the high-resolution three-dimensional image based on a transfer learning model to generate dynamic texture features associated with the blood drying time.

[0060] Specifically, select a deep convolutional neural network (CNN) model pre-trained on a large-scale image dataset (such as ImageNet), such as ResNet or VGG, as the base model for transfer learning. Freeze some layers of the pre-trained model and only fine-tune some layers to adapt to the specific features of bloodstain images. Use the convolutional layers of the transfer learning model to extract the depth features of high-resolution three-dimensional images, and the above features can capture the microscopic texture and structure of bloodstains. Input the time series information of bloodstain images (such as images at different drying stages) into the model to capture the changes of dynamic texture features over time. Through time series analysis or regression models, associate the extracted dynamic texture features with the blood drying time to generate time-related feature vectors. Verify through comparative experiments whether the extracted dynamic texture features can accurately reflect the changes during the blood drying process. Adjust the model parameters according to the verification results to optimize the feature extraction effect.

[0061] Through the above steps, it is possible to use the transfer learning model to extract the dynamic texture features of bloodstain images and associate them with the blood drying time, providing high-quality feature inputs for subsequent time estimation and parameter correction.

[0062] The time texture modeling module 103 is used to perform multimodal fusion processing by combining environmental parameters and dynamic texture features, construct a time-texture mapping library, and generate a probability distribution model of the bloodstain formation time.

[0063] Specifically, collect temperature and humidity parameter data of the environment where the target bloodstain is located to generate environmental parameter data. Perform feature vector splicing on the dynamic texture features and environmental parameter data to generate a fused feature vector, realizing the fusion of multimodal data. Label the drying time tags of historical bloodstain samples, and synchronously collect the corresponding historical dynamic texture features and historical environmental parameter data. Perform noise reduction processing on the historical dynamic texture features to generate standardized dynamic texture features and improve the data quality. Perform normalization processing on the historical environmental parameter data and standardized dynamic texture features to generate a standardized time-texture dataset to ensure the consistency and comparability of the data. Associate and store the standardized time-texture dataset with the drying time tags to construct a time-texture mapping library, providing data support for subsequent model training. Based on the time-texture mapping library, train a probability distribution model of the bloodstain formation time through a Bayesian regression model. This model can take into account the influence of various uncertain factors and give a probability distribution within a time range. Through the above steps, it is possible to make full use of environmental parameters and dynamic texture features to construct a probability distribution model that accurately reflects the bloodstain formation time, providing a more reliable basis for bloodstain analysis.

[0064] The time estimation module 104 is used to calculate the formation time according to the probability distribution model, generate the formation time of the target bloodstain, and correct the calculation results of the dripping height and angle parameters based on the formation time to obtain the corrected dripping height and angle parameters.

[0065] Specifically, using the pre-constructed probability distribution model of bloodstain formation time, input the relevant characteristic data of the target bloodstain, including dynamic texture features and environmental parameters, etc., to calculate the probability distribution of bloodstain formation time. Extract the dry time probability distribution data of the target bloodstain from the probability distribution model, generate the time estimation value and confidence interval parameters, and determine the most likely range of bloodstain formation time. Based on the high-resolution three-dimensional image, calculate the initial dripping height parameter and initial contact angle parameter through the hydrodynamics model, providing a basis for subsequent correction. According to the time estimation value, perform time decay compensation calculation on the initial dripping height parameter and initial contact angle parameter to generate the corrected dripping height parameter and contact angle parameter. This step considers the variation law of blood viscosity with dry time, and adjusts the initial parameters by establishing the mapping relationship between time-viscosity decay coefficient.

[0066] Based on the confidence interval parameters, verify the error boundary of the corrected dripping height parameter and contact angle parameter through Monte Carlo simulation, generate the corrected dripping height and angle parameters, and ensure the accuracy and reliability of the parameters. Perform hydrodynamic simulation verification on the corrected dripping height parameter and contact angle parameter to ensure that it meets the physical constraint conditions of bloodstain diffusion morphology, further improving the accuracy of the parameters. Through the above steps, it is possible to accurately calculate the bloodstain formation time based on the probability distribution model, effectively correct the dripping parameters, and improve the accuracy and reliability of bloodstain analysis.

[0067] The three-dimensional reconstruction module 105 is used to reconstruct the three-dimensional bleeding point position based on the corrected dripping height and angle parameters through the back-projection algorithm.

[0068] Specifically, based on the spatial information of the scene where the bloodstain is located, a three-dimensional spatial coordinate system is established to provide a reference framework for subsequent reverse projection calculations. The corrected drop height parameters and contact angle parameters are input into the reverse projection algorithm to generate the initial projection parameters of the bloodstain motion trajectory. The above parameters describe the motion trajectory of the bloodstain from the bleeding point to the target surface. Using the initial projection parameters and the fluid mechanics model, the bloodstain motion trajectory is reversely projected to generate a three-dimensional spatial trajectory path. This step simulates the motion process of the bloodstain through a mathematical model to infer the original motion trajectory of the bloodstain. The three-dimensional coordinates of the bleeding point are determined according to the intersection of the three-dimensional spatial trajectory path and the physical constraints of the scene. Through geometric collision detection, the intersection that meets the bloodstain diffusion speed and direction constraints is selected. Verify whether the three-dimensional coordinates conform to the fluid dynamics law of the bloodstain diffusion morphology. If not, iteratively adjust the projection parameters to ensure that the reconstructed three-dimensional bleeding point position is accurate and reliable. Through the above steps, the three-dimensional bleeding point position can be accurately reconstructed using the reverse projection algorithm and the corrected drop parameters, providing key clues for case analysis.

[0069] An automatic bloodstain morphology analysis and drip parameter calculation system based on machine vision provided by an embodiment of the present application includes: a multimodal image acquisition module, which is used to perform multimodal image acquisition on a target bloodstain to obtain a high-resolution three-dimensional image including a microscopic crystal structure; a dynamic texture feature extraction module, which is used to perform dynamic texture feature extraction on the high-resolution three-dimensional image based on a transfer learning model to generate dynamic texture features associated with the blood drying time; a time texture modeling module, which is used to perform multimodal fusion processing in combination with environmental parameters and dynamic texture features, build a time-texture mapping library, and generate a probability distribution model of bloodstain formation time; a time estimation module, which is used to calculate the formation time according to the probability distribution model, generate the formation time of the target bloodstain, and correct the parameter calculation results of the drip height and angle based on the formation time to obtain the corrected drip height and angle parameters; a three-dimensional reconstruction module, which is used to reconstruct the three-dimensional bleeding point position through a reverse projection algorithm based on the corrected drip height and angle parameters, so as to solve the limitations of the relevant bloodstain analysis technology in time estimation and parameter correction, thereby improving the accuracy and scene adaptability of the bloodstain analysis.

[0070] In one embodiment, if Figure 2 As shown, the formation time is calculated according to the probability distribution model to generate the formation time of the target bloodstain, and the parameter calculation results of the drop height and angle are corrected based on the formation time to obtain the corrected drop height and angle parameters, including:

[0071] S201: extracting drying time probability distribution data of the target bloodstain from the probability distribution model, and generating a time estimation value and a confidence interval parameter;

[0072] S202: Calculate the initial dripping height parameter and the initial contact angle parameter based on the high-resolution three-dimensional image through a hydrodynamic model;

[0073] S203: Perform time decay compensation calculations on the initial dripping height parameter and the initial contact angle parameter according to the time estimate value to generate the corrected dripping height parameter and the contact angle parameter;

[0074] S204: Verify the error boundaries of the corrected dripping height parameter and the contact angle parameter through Monte Carlo simulation based on the confidence interval parameter to generate the corrected dripping height and angle parameters.

[0075] Specifically, extract the dry time probability distribution data of the target bloodstain from the probability distribution model to generate the time estimate value and the confidence interval parameter. This step provides a time basis for subsequent parameter correction. Calculate the initial dripping height parameter and the initial contact angle parameter based on the high-resolution three-dimensional image through a hydrodynamic model. This step uses image data and physical models to provide basic parameters for subsequent correction. Perform time decay compensation calculations on the initial dripping height parameter and the initial contact angle parameter according to the time estimate value to generate the corrected dripping height parameter and the contact angle parameter. This step considers the variation law of blood viscosity with the dry time, and adjusts the initial parameters by establishing a mapping relationship between time and viscosity decay coefficient. Verify the error boundaries of the corrected dripping height parameter and the contact angle parameter through Monte Carlo simulation based on the confidence interval parameter to generate the corrected dripping height and angle parameters. This step ensures that the corrected parameters are within a reasonable error range through simulation and verification. Through the above steps, it is possible to accurately calculate the bloodstain formation time based on the probability distribution model, effectively correct the dripping parameters, and improve the accuracy and reliability of bloodstain analysis.

[0076] Further, performing time decay compensation calculations on the initial dripping height parameter and the initial contact angle parameter according to the time estimate value to generate the corrected dripping height parameter and the contact angle parameter includes:

[0077] Based on the variation law of blood viscosity with the dry time, establish a mapping relationship between time and viscosity decay coefficient;

[0078] According to the time estimate value, match the corresponding viscosity decay coefficient from the mapping relationship;

[0079] Based on the viscosity decay coefficient, perform time-varying viscosity compensation calculations on the initial dripping height parameter to generate the corrected dripping height parameter;

[0080] Based on the viscosity decay coefficient, perform time-varying viscosity compensation calculations on the initial contact angle parameter to generate the corrected contact angle parameter;

[0081] Perform hydrodynamic simulation verification on the corrected droplet height parameter and contact angle parameter to ensure that they meet the physical constraints of the bloodstain diffusion pattern.

[0082] Specifically, extract the dry time probability distribution data of the target bloodstain from the probability distribution model to generate time estimate values and confidence interval parameters. Based on the variation law of blood viscosity with dry time, establish a mapping relationship between time and viscosity attenuation coefficient. This step uses rheology theory and experimental data to ensure the accuracy of the mapping relationship. According to the time estimate values, match the corresponding viscosity attenuation coefficients from the mapping relationship to provide a basis for subsequent compensation calculations. Based on the viscosity attenuation coefficients, perform time-varying viscosity compensation calculations on the initial droplet height parameter and initial contact angle parameter respectively to generate the corrected droplet height parameter and contact angle parameter. Perform hydrodynamic simulation verification on the corrected parameters to ensure that they meet the physical constraints of the bloodstain diffusion pattern. This step verifies the rationality of the corrected parameters through numerical simulation to ensure the reliability of the analysis results.

[0083] Furthermore, based on the confidence interval parameters, perform error boundary verification on the corrected droplet height parameter and contact angle parameter through Monte Carlo simulation to generate the corrected droplet height and angle parameters, including:

[0084] Generate random parameter combinations of droplet height and contact angle based on the confidence interval parameters;

[0085] Perform hydrodynamic simulation on the random parameter combinations to generate bloodstain diffusion pattern simulation results;

[0086] Based on the comparison between the bloodstain diffusion pattern simulation results and the actual shape data of the target bloodstain, calculate the parameter error distribution;

[0087] Determine the error boundaries of the corrected droplet height parameter and contact angle parameter according to the parameter error distribution, and generate the corrected droplet height and angle parameters based on the error boundaries.

[0088] Specifically, based on the confidence interval parameters, random parameter combinations of the dripping height and contact angle are generated. This step utilizes the random sampling characteristic of Monte Carlo simulation to generate multiple possible parameter combinations, ensuring coverage of a reasonable range of parameters. Hydrodynamic simulations are performed on the generated random parameter combinations to simulate the bloodstain diffusion pattern. Through numerical simulation, simulation results of the bloodstain diffusion pattern are generated, providing data support for subsequent error analysis. The simulation results of the bloodstain diffusion pattern are compared with the actual morphological data of the target bloodstain, and the parameter error distribution is calculated. This step evaluates the difference between the simulation results and the actual data through statistical analysis to determine the distribution characteristics of the errors. Based on the parameter error distribution, the error boundaries of the corrected dripping height parameter and contact angle parameter are determined. This step ensures that the corrected parameters are within a reasonable error range through error analysis, improving the reliability and accuracy of the parameters. Through the above steps, it is possible to use Monte Carlo simulation and hydrodynamic simulation to verify the error boundaries of the corrected dripping height and angle parameters, ensuring the accuracy and reliability of the analysis results.

[0089] Furthermore, based on the corrected dripping height and angle parameters, the three-dimensional bleeding point position is reconstructed through the back-projection algorithm, including:

[0090] Based on the spatial information of the scene where the bloodstain is located, a three-dimensional space coordinate system is established;

[0091] The corrected dripping height parameter and contact angle parameter are input into the back-projection algorithm to generate the initial projection parameters of the bloodstain movement trajectory;

[0092] Based on the initial projection parameters and the hydrodynamic model, reverse projection calculation is performed on the bloodstain movement trajectory to generate a three-dimensional space trajectory path;

[0093] According to the intersection point of the three-dimensional space trajectory path and the scene physical constraint conditions, the three-dimensional coordinates of the bleeding point are determined;

[0094] Verify whether the three-dimensional coordinates conform to the hydrodynamic laws of the bloodstain diffusion pattern. If not, iterate and adjust the projection parameters.

[0095] Specifically, based on the spatial information of the scene where the bloodstain is located, a three-dimensional space coordinate system is established to provide a reference framework for subsequent back-projection calculations. The corrected dripping height parameter and contact angle parameter are input into the reverse projection algorithm to generate the initial projection parameters of the bloodstain movement trajectory. The above parameters describe the movement trajectory of the bloodstain from the bleeding point to the target surface. Using the initial projection parameters and the hydrodynamic model, a back-projection calculation is performed on the bloodstain movement trajectory to generate a three-dimensional space trajectory path. This step simulates the movement process of the bloodstain through a mathematical model and calculates the original movement trajectory of the bloodstain. According to the intersection points of the three-dimensional space trajectory path and the scene physical constraint conditions, the three-dimensional coordinates of the bleeding point are determined. Through geometric collision detection, the intersection points that meet the bloodstain diffusion speed and direction constraints are screened out. Verify whether the three-dimensional coordinates conform to the hydrodynamic laws of the bloodstain diffusion pattern. If not, the projection parameters are iteratively adjusted to ensure the accuracy and reliability of the reconstructed three-dimensional bleeding point position.

[0096] Furthermore, according to the intersection points of the three-dimensional space trajectory path and the scene physical constraint conditions, determining the three-dimensional coordinates of the bleeding point includes:

[0097] Perform three-dimensional modeling on the physical boundary conditions of the scene where the bloodstain is located to generate a scene constraint model;

[0098] Based on the three-dimensional space trajectory path and the scene constraint model, perform geometric collision detection to generate a set of candidate intersection points;

[0099] Verify the hydrodynamic laws for the set of candidate intersection points and screen out the intersection points that meet the bloodstain diffusion speed and direction constraints;

[0100] Based on the screened intersection points, determine the three-dimensional coordinates of the optimal bleeding point through a spatial clustering algorithm.

[0101] Specifically, perform three-dimensional modeling on the physical boundary conditions of the scene where the bloodstain is located to generate a scene constraint model. This step digitizes the physical characteristics of the scene to provide a basis for subsequent collision detection. Based on the three-dimensional space trajectory path and the scene constraint model, perform geometric collision detection to generate a set of candidate intersection points. By calculating the intersection points of the trajectory path and the scene model, possible bleeding point positions are screened out. Verify the hydrodynamic laws for the set of candidate intersection points and screen out the intersection points that meet the bloodstain diffusion speed and direction constraints. This step ensures the physical reasonableness of the candidate intersection points and excludes the points that do not conform to the bloodstain diffusion law. Based on the screened intersection points, determine the three-dimensional coordinates of the optimal bleeding point through a spatial clustering algorithm. The spatial clustering algorithm can find the most likely bleeding point position according to the spatial distribution characteristics of the intersection points, improving the accuracy of positioning.

[0102] Furthermore, perform multi-modal fusion processing by combining environmental parameters and dynamic texture features to construct a time-texture mapping library and generate a probability distribution model of the bloodstain formation time, including:

[0103] Collect the temperature and humidity parameters of the environment where the target bloodstain is located to generate environmental parameter data;

[0104] Perform feature vector splicing processing on the dynamic texture features and environmental parameter data to generate a fused feature vector;

[0105] Construct a time-texture mapping library based on the drying time labels of historical bloodstain samples, historical dynamic texture features, and historical environmental parameter data;

[0106] Based on the time-texture mapping library, generate a probability distribution model of the bloodstain formation time through Bayesian regression model training.

[0107] Specifically, collect the temperature and humidity parameters of the environment where the target bloodstain is located to generate environmental parameter data. Perform feature vector splicing processing on the dynamic texture features and environmental parameter data to generate a fused feature vector. This step integrates information of different modalities into a stable multi-modal representation through feature-level fusion. Label the drying time labels of historical bloodstain samples, and synchronously collect the corresponding historical dynamic texture features and historical environmental parameter data. Perform noise reduction processing on the historical dynamic texture features to generate standardized dynamic texture features and improve the data quality. Perform normalization processing on the historical environmental parameter data and standardized dynamic texture features to generate a standardized time-texture data set to ensure the consistency and comparability of the data. Associate and store the standardized time-texture data set with the drying time labels to construct a time-texture mapping library to provide data support for subsequent model training. Based on the time-texture mapping library, generate a probability distribution model of the bloodstain formation time through Bayesian regression model training. The Bayesian regression model expresses the uncertainty of parameters by introducing a prior distribution and uses the observed data to update the posterior distribution of the parameters, thereby generating a probability distribution of the bloodstain formation time. Through the above steps, it is possible to make full use of environmental parameters and dynamic texture features to construct a probability distribution model that accurately reflects the bloodstain formation time and provide a more reliable basis for bloodstain analysis.

[0108] Furthermore, construct a time-texture mapping library based on the drying time labels of historical bloodstain samples, historical dynamic texture features, and historical environmental parameter data, including:

[0109] Label the drying time labels of historical bloodstain samples, and synchronously collect the corresponding historical dynamic texture features and historical environmental parameter data;

[0110] Perform noise reduction processing on the historical dynamic texture features to generate standardized dynamic texture features;

[0111] Perform normalization processing on the historical environmental parameter data and standardized dynamic texture features to generate a standardized time-texture data set;

[0112] Associate and store the standardized time-texture dataset with the drying time tags to construct a time-texture mapping library.

[0113] Specifically, accurately label the drying time of historical bloodstain samples to form drying time tags. At the same time, synchronously collect the historical dynamic texture features and historical environmental parameter data of these samples. Denoise the collected historical dynamic texture features to remove noise interference and generate clearer and more standardized dynamic texture features. Normalize the historical environmental parameter data and the standardized dynamic texture features to make the data in different dimensions comparable and generate a standardized time-texture dataset. Associate and store the standardized dynamic texture features and environmental parameter data with the corresponding drying time tags to construct a time-texture mapping library. This step provides structured data support for subsequent model training by establishing the association relationships between the data.

[0114] In one embodiment, by loading a bloodstain image and applying a feature extraction method to obtain the key features of the bloodstain, combined with standardization processing, principal component analysis (PCA), and a linear regression model, analyze the bloodstain features to predict the bloodstain type and related parameters. By integrating a machine learning model into bloodstain analysis, a highly automated analysis process is achieved, reducing human errors. The custom feature extraction method combined with PCA dimensionality reduction technology significantly improves the analysis accuracy. Different analysis logics are developed for dynamic and static bloodstains to support the application of multiple crime scenes. By designing an effective feature extraction algorithm, capture the key features of the bloodstain morphology, and use the linear regression model and PCA technology to accurately predict the bloodstain type and estimate the related parameters. Standardize the features through StandardScaler to ensure the unity of the model input data, and make the accuracy rate of bloodstain type classification and related parameter prediction reach more than 90%. The optimized feature extraction algorithm controls the feature extraction time of a single image within a short time, supports multiple image formats such as.jpg and.png, and is applicable to the analysis of different types of bloodstains.

[0115] In one embodiment, by designing a dedicated feature extraction algorithm and combining it with principal component analysis (PCA), bloodstain features are extracted from multiple dimensions, significantly improving the accuracy and scientific nature of the analysis. A linear regression model is used to predict bloodstain-related parameters (such as the dropping height), providing a scientific basis for crime scene reconstruction. Dynamic and static bloodstains are distinguished: for static bloodstains, the dropping height and contact angle are calculated through the regression model, and three-dimensional reconstruction information is provided in combination with the feature extraction results; for dynamic bloodstains, the motion intensity index is calculated based on the bloodstain features to infer the speed and state of the perpetrator. In addition, StandardScaler is used to standardize the features, ensuring the unity of the model input data, enhancing the generalization ability of the model, and a modular design is adopted to support future extended analysis of more bloodstain types.

[0116] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the indications of the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear description in this article, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same moment, but can be executed at different moments. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0117] In one embodiment, the present application also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps in the above-described system embodiments are implemented.

[0118] In one embodiment, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-described system embodiments are implemented.

[0119] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can refer to the partial descriptions of the method embodiments. The device embodiments described above are only illustrative. The components described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the present disclosure solution. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0120] The above-described embodiments merely represent several implementation manners of the embodiments of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent of the embodiments of the application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the embodiments of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the embodiments of the present application.

Claims

1. An automatic analysis system for bloodstain patterns and calculation system for dripping parameters based on machine vision, characterized in that, The system includes: A multi-modal image acquisition module, which is used to perform multi-modal image acquisition on the target blood stain to obtain a high-resolution three-dimensional image including a microscopic crystal structure; A dynamic texture feature extraction module, which is used to extract dynamic texture features from the high-resolution three-dimensional image based on a transfer learning model to generate dynamic texture features associated with the blood drying time; A time-texture modeling module, which is used to perform multi-modal fusion processing by combining environmental parameters and the dynamic texture features, construct a time-texture mapping library, and generate a probability distribution model of the blood stain formation time; A time estimation module, which is used to calculate the formation time according to the probability distribution model to generate the formation time of the target blood stain, and correct the calculation results of the dripping height and angle parameters based on the formation time to obtain the corrected dripping height and angle parameters; A three-dimensional reconstruction module, which is used to reconstruct the three-dimensional bleeding point position based on the corrected dripping height and angle parameters through a back-projection algorithm.

2. The system for automatically analyzing bloodstain patterns and calculating dripping parameters based on machine vision according to claim 1, wherein The calculating the formation time according to the probability distribution model to generate the formation time of the target blood stain, and correcting the calculation results of the dripping height and angle parameters based on the formation time to obtain the corrected dripping height and angle parameters includes: Extracting the dry time probability distribution data of the target blood stain from the probability distribution model to generate a time estimation value and confidence interval parameters; Based on the high-resolution three-dimensional image, calculating the initial dripping height parameter and the initial contact angle parameter through a hydrodynamic model; According to the time estimation value, performing time decay compensation calculation on the initial dripping height parameter and the initial contact angle parameter to generate corrected dripping height parameters and contact angle parameters; Based on the confidence interval parameters, verifying the error boundary of the corrected dripping height parameters and contact angle parameters through Monte Carlo simulation to generate the corrected dripping height and angle parameters.

3. The automatic bloodstain pattern analysis and dripping parameter calculation system based on machine vision according to claim 2, wherein The performing time decay compensation calculation on the initial dripping height parameter and the initial contact angle parameter according to the time estimation value to generate corrected dripping height parameters and contact angle parameters includes: Establishing a mapping relationship between time and viscosity decay coefficient based on the variation law of blood viscosity with drying time; According to the time estimation value, matching the corresponding viscosity decay coefficient from the mapping relationship; Based on the viscosity decay coefficient, performing time-varying viscosity compensation calculation on the initial dripping height parameter to generate the corrected dripping height parameter; Based on the viscosity decay coefficient, performing time-varying viscosity compensation calculation on the initial contact angle parameter to generate the corrected contact angle parameter; Performing hydrodynamic simulation verification on the corrected dripping height parameters and contact angle parameters to ensure that they meet the physical constraint conditions of the blood stain diffusion pattern.

4. The automatic analysis system for bloodstain pattern and calculation system for dripping parameters based on machine vision according to claim 2, wherein The verifying the error boundary of the corrected dripping height parameters and contact angle parameters through Monte Carlo simulation based on the confidence interval parameters to generate the corrected dripping height and angle parameters includes: Generating a random parameter combination of dripping height and contact angle based on the confidence interval parameters; Perform a hydrodynamic simulation on the combination of random parameters to generate a simulation result of the bloodstain diffusion pattern; Based on the comparison between the simulation result of the bloodstain diffusion pattern and the actual morphological data of the target bloodstain, calculate the parameter error distribution; Determine the error boundaries of the corrected dropping height parameter and contact angle parameter according to the parameter error distribution, and generate the corrected dropping height and angle parameters based on the error boundaries.

5. The bloodstain pattern automatic analysis and dripping parameter calculation system based on machine vision according to claim 1, characterized in that, Based on the corrected dropping height and angle parameters, reconstruct the three-dimensional bleeding point position through the back-projection algorithm, including: Establish a three-dimensional space coordinate system based on the spatial information of the scene where the bloodstain is located; Input the corrected dropping height parameter and contact angle parameter into the back-projection algorithm to generate the initial projection parameters of the bloodstain movement trajectory; Based on the initial projection parameters and the hydrodynamics model, perform back-projection calculation on the bloodstain movement trajectory to generate a three-dimensional space trajectory path; Determine the three-dimensional coordinates of the bleeding point according to the intersection points of the three-dimensional space trajectory path and the scene physical constraint conditions; Verify whether the three-dimensional coordinates conform to the hydrodynamic law of the bloodstain diffusion pattern, and if not, iteratively adjust the projection parameters.

6. The automatic analysis system for bloodstain morphology and calculation system for dripping parameters based on machine vision according to claim 5, characterized in that The determining the three-dimensional coordinates of the bleeding point according to the intersection points of the three-dimensional space trajectory path and the scene physical constraint conditions includes: Perform three-dimensional modeling on the physical boundary conditions of the scene where the bloodstain is located to generate a scene constraint model; Perform geometric collision detection based on the three-dimensional space trajectory path and the scene constraint model to generate a set of candidate intersection points; Verify the hydrodynamic law of the set of candidate intersection points, and screen out the intersection points that meet the bloodstain diffusion speed and direction constraints; Based on the screened intersection points, determine the optimal three-dimensional coordinates of the bleeding point through a spatial clustering algorithm.

7. The automatic analysis system for bloodstain morphology and calculation system for dripping parameters based on machine vision according to claim 1, wherein The combining the environmental parameters and the dynamic texture features for multimodal fusion processing to construct a time-texture mapping library and generate a probability distribution model of the bloodstain formation time includes: Collect the temperature and humidity parameter of the environment where the target bloodstain is located to generate environmental parameter data; Perform feature vector splicing processing on the dynamic texture features and the environmental parameter data to generate a fused feature vector; Construct a time-texture mapping library based on the dry time labels of historical bloodstain samples, historical dynamic texture features, and historical environmental parameter data; Based on the time-texture mapping library, train a probability distribution model of the bloodstain formation time through a Bayesian regression model.

8. The automatic analysis system of bloodstain morphology and calculation system of dripping parameters based on machine vision according to claim 7, characterized in that, The constructing a time-texture mapping library based on the dry time labels of historical bloodstain samples, historical dynamic texture features, and historical environmental parameter data includes: Annotate the dry time labels of the historical bloodstain samples, and synchronously collect the corresponding historical dynamic texture features and historical environmental parameter data; Perform noise reduction processing on the historical dynamic texture features to generate standardized dynamic texture features; Perform normalization processing on the historical environmental parameter data and the standardized dynamic texture features to generate a standardized time-texture data set; Associate and store the standardized time-texture data set with the dry time labels to construct the time-texture mapping library.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the machine vision-based automatic bloodstain pattern analysis and dripping parameter calculation system according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the machine vision-based automatic bloodstain pattern analysis and dripping parameter calculation system according to any one of claims 1 to 8.

Images