Optimal view angle selection method and device for three-dimensional property right body, equipment and medium

Abstract

The invention relates to the technical field of computer vision, and discloses an optimal view angle selection method and device for a three-dimensional property right body, equipment and a medium, and the method comprises the steps: obtaining a feature parameter set and a score label corresponding to each candidate view angle under a plurality of three-dimensional property right bodies, and constructing a training data set; training a support vector machine based on a genetic algorithm based on the training data set, and determining the optimal evaluation weight of each feature parameter relative to the visual score; obtaining a target feature parameter set corresponding to each candidate view angle of the target three-dimensional property right body, and determining a view angle score corresponding to each candidate view angle in combination with the optimal evaluation weight; according to the method, the weights of the characteristic parameters are determined through the genetic algorithm and the support vector machine, so that score calculation is carried out to determine the optimal view angle of the target three-dimensional property right body, information can be obtained as much as possible under the optimal view angle, and the accuracy of the target three-dimensional property right body is improved. And user experience is improved.

Description

Technical Field

[0001] The present invention relates to the field of computer vision technology, and in particular to a method, device, equipment and medium for selecting an optimal viewing angle of a three-dimensional property. Background Art

[0002] The three-dimensional cadastral technology realizes the intuitive spatial display of the three-dimensional cadastral through three-dimensional modeling, digitization, virtual reality, augmented reality and other technologies, which makes up for the deficiency of the two-dimensional cadastral technology in fully and accurately reflecting the characteristics of geographic information. However, due to the diversity and complexity of real estate, its characteristics cannot be fully presented through a single perspective. The traditional three-dimensional cadastral mapping technology lacks the ability to fully display the multi-dimensional information of real estate, and cannot accurately display the characteristics of complex terrain, special buildings and large-scale real estate. The management and transaction of real estate rights need to be based on accurate and reliable cadastral information support. Through the optimal perspective selection technology, the geographic information characteristics of real estate can be better presented, providing an important basis for the confirmation and transaction of rights.

[0003] Currently, when displaying three-dimensional cadastral property rights, an optimal viewing angle is often selected manually based on experience for observation. Under the optimal viewing angle determined in this way, users often cannot obtain as much visual information as possible, which affects the user experience. Summary of the invention

[0004] In view of this, the present invention provides a method, device, computer equipment and storage medium for selecting the best viewing angle of a three-dimensional property object to solve the problem that users cannot obtain as much visual information as possible and the user experience is poor.

[0005] In a first aspect, the present invention provides a method for selecting an optimal viewing angle of a three-dimensional property, the method comprising:

[0006] Obtaining feature parameter sets and corresponding scoring labels corresponding to each candidate perspective under multiple three-dimensional property bodies;

[0007] The feature parameter set and the corresponding score label corresponding to each candidate perspective are used as a piece of training data to construct a training data set;

[0008] Training a support vector machine based on a genetic algorithm based on the training data set to determine an optimal evaluation weight of each feature parameter in the feature parameter set relative to a visual score;

[0009] Obtaining a target feature parameter set corresponding to each candidate viewing angle of the target three-dimensional property, and determining a viewing angle score corresponding to each candidate viewing angle in combination with an optimal evaluation weight of each feature parameter relative to a visual score;

[0010] The optimal viewing angle of the target three-dimensional property object is determined according to the viewing angle scores corresponding to each candidate viewing angle.

[0011] The method provided in this aspect trains a support vector machine based on a genetic algorithm through training data with corresponding score labels, so as to determine the optimal evaluation weights corresponding to various feature parameters that can affect the perspective score, and then for a target three-dimensional property body that needs to select the best perspective, the feature parameter set of the target three-dimensional property body at each perspective and the optimal evaluation weights corresponding to each feature parameter are calculated, and the scores of each perspective are determined to select the best perspective. This can ensure the scientific nature of the selection of the best perspective, improve the user's information acquisition at the best perspective, and ensure user experience.

[0012] In an optional implementation, the training of a support vector machine based on a genetic algorithm based on the training data set to determine the optimal evaluation weight of each feature parameter in the feature parameter set relative to the visual score includes:

[0013] Determine an initialization population corresponding to the genetic algorithm, and iteratively optimize the initialization population, wherein each individual in the initialization population corresponds to a weight group;

[0014] In each iterative optimization process, the fitness of each individual in the population is determined based on the training data set and the support vector machine, and the population is iterated through a genetic algorithm based on the fitness of each individual in the population;

[0015] When the iteration termination condition of the genetic algorithm is reached, multiple qualified weight groups are determined according to the fitness of each individual in the current population;

[0016] The support vector machine is trained based on the multiple qualified weight groups and the training data to determine the optimal evaluation weight of each feature parameter in the feature parameter set relative to the visual score.

[0017] In this implementation mode, the initialization population of the genetic algorithm is continuously iterated and optimized, and the individual fitness is determined by a support vector machine in each iterative optimization to ensure the iterative effect. After the iteration of the genetic algorithm is completed, the support vector machine is further trained based on multiple qualified weight groups to finally determine the optimal evaluation weights for each feature parameter, thereby ensuring that the ultimately determined optimal evaluation weights can effectively characterize the impact of each feature parameter on the score, thereby improving the accuracy of subsequent score calculations.

[0018] In an optional implementation, during each iterative optimization process, determining the fitness of each individual in the population based on the training data set and the support vector machine includes:

[0019] For each individual in the population, the weight group corresponding to the individual is substituted into the support vector machine, and each training data in the training data set is predicted to obtain the prediction score corresponding to each training data;

[0020] The fitness of the individual is calculated according to the predicted scores and score labels corresponding to the respective training data.

[0021] In this implementation, the fitness of each individual is determined by a support vector machine, which can effectively ensure the scientific nature of fitness calculation and improve the iterative effect of subsequent genetic algorithms.

[0022] In an optional implementation manner, the candidate viewing angles are determined as follows:

[0023] For each three-dimensional property body, based on the center position of the three-dimensional property body and the center position of the sphere under the preset radius, the three-dimensional property body is set in the sphere under the preset radius, and the center position of the three-dimensional property body and the center position of the sphere coincide with each other;

[0024] Constructing a spherical coordinate system of a sphere corresponding to the three-dimensional property body, and determining a plurality of grid points on the upper hemisphere of the sphere according to a preset polar angle range and a preset azimuth angle range;

[0025] For each grid point, the viewing angle of the grid point toward the sphere center is determined as a candidate viewing angle.

[0026] In this embodiment, the surface of the sphere is divided within a preset polar angle range and azimuth angle range in the spherical coordinate system corresponding to the three-dimensional property body, thereby obtaining multiple grid points, and then determining multiple candidate viewing angles, which can effectively ensure the uniform arrangement of the candidate viewing angles and ensure the scientific nature of the arrangement of the candidate viewing angles.

[0027] In an optional implementation, the preset radius is determined as follows:

[0028] For each three-dimensional property volume, determining a three-dimensional bounding box of the three-dimensional property volume;

[0029] The preset radius of the sphere corresponding to the three-dimensional property body is determined based on the distance between each boundary point in the three-dimensional boundary frame and the center position of the three-dimensional property body.

[0030] In this embodiment, the preset radius of the sphere corresponding to the three-dimensional property body is determined by the three-dimensional bounding box of the three-dimensional property body, which can effectively ensure that each subsequent viewing angle of the surface of the sphere can have a relatively reasonable visual range.

[0031] In an optional embodiment, the characteristic parameter set includes: visibility, visible area ratio, surface area entropy, visual connectivity, viewing angle comfort, visual coherence and viewing angle variability;

[0032] Determining the optimal evaluation weight of each feature parameter in the feature parameter set relative to the visual score comprises:

[0033] The optimal evaluation weights of visibility, visible area ratio, surface area entropy, visual connectivity, viewing angle comfort, visual coherence and viewing angle variability in the feature parameter set relative to the visual score are determined.

[0034] In this implementation, by determining the optimal evaluation weights of multiple different feature parameters under the viewing angle, it can be ensured that when the viewing angle is subsequently scored, the effect of the viewing angle can be evaluated as reasonably as possible from multiple dimensions.

[0035] In an optional implementation, for each candidate viewing angle under each three-dimensional property body, the method for obtaining each feature parameter in the feature parameter set corresponding to the candidate viewing angle is as follows:

[0036] Determining the visibility of the candidate viewing angle according to the number of visible three-dimensional boundary points under the candidate viewing angle and the total number of boundary points of the three-dimensional property body;

[0037] Determining a visible area ratio of the candidate viewing angle according to a projection area of ​​the candidate viewing angle relative to the three-dimensional property body and a surface area of ​​the three-dimensional property body;

[0038] Determining the surface area entropy at the candidate viewing angle according to the area of ​​each visible boundary surface at the candidate viewing angle;

[0039] Determine the connectivity of the sight line under the candidate viewing angle according to the length of the connected covered sight line under the candidate viewing angle and the total length of the parcel body of the three-dimensional property body;

[0040] Determining the viewing comfort of the candidate viewing angle according to the angles between the candidate viewing angle and the most comfortable viewing angle of the three-dimensional object relative to the longitudinal axis of the three-dimensional object;

[0041] Determining visual coherence of the candidate perspectives according to the number of consistent features of the candidate perspectives and the total number of features;

[0042] The visual variability of the candidate viewing angle is determined based on the degree of variability between the candidate viewing angle and other candidate viewing angles in the three-dimensional property.

[0043] In this implementation manner, the feature parameters corresponding to each viewing angle are determined by using the relevant parameters of the three-dimensional property under the candidate viewing angles, so as to ensure the validity of the subsequent viewing angle scoring.

[0044] In a second aspect, the present invention provides a device for selecting an optimal viewing angle of a three-dimensional property, the device comprising:

[0045] A data acquisition module, used to obtain a feature parameter set and a corresponding scoring label corresponding to each candidate perspective under multiple three-dimensional property bodies;

[0046] A training set construction module is used to construct a training data set by taking the feature parameter set and the corresponding score label corresponding to each candidate perspective as a piece of training data;

[0047] A weight determination module, used for training a support vector machine based on a genetic algorithm based on the training data set, and determining an optimal evaluation weight of each feature parameter in the feature parameter set relative to a visual score;

[0048] A viewing angle scoring module is used to obtain a target feature parameter set corresponding to each candidate viewing angle of the target three-dimensional property object, and determine a viewing angle score corresponding to each candidate viewing angle in combination with the best evaluation weight of each feature parameter relative to the visual score;

[0049] The viewing angle determination module is used to determine the best viewing angle of the target three-dimensional property object according to the viewing angle scores corresponding to each candidate viewing angle.

[0050] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the method for selecting an optimal viewing angle of a three-dimensional property body according to the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0051] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the method for selecting an optimal viewing angle of a three-dimensional property body according to the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0053] Figure 1 is a schematic flow chart of a method for selecting an optimal viewing angle of a three-dimensional property according to an embodiment of the present invention;

[0054] Figure 2 is a flow chart of another method for selecting an optimal viewing angle of a three-dimensional property according to an embodiment of the present invention;

[0055] Figure 3is an example diagram of viewpoint deployment of a three-dimensional property according to an embodiment of the present invention;

[0056] Figure 4 It is a schematic diagram of observing a point target, a line target, and a surface target under a three-dimensional boundary according to an embodiment of the present invention;

[0057] Figure 5 An example diagram of an implementation process for determining an optimal viewing angle of a three-dimensional cadastral property body is provided according to an embodiment of the present invention;

[0058] Figure 6 An exemplary diagram of an implementation process for determining the optimal viewing angle of a three-dimensional cadastral property body is provided according to an embodiment of the present invention.

[0059] Figure 7 is a structural block diagram of an optimal viewing angle selection device for a three-dimensional property according to an embodiment of the present invention;

[0060] Figure 8 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0062] The three-dimensional cadastral technology realizes the intuitive spatial display of the three-dimensional cadastral through three-dimensional modeling, digitization, virtual reality, augmented reality and other technologies, which makes up for the deficiency of the two-dimensional cadastral technology in fully and accurately reflecting the characteristics of geographic information. However, due to the diversity and complexity of real estate, its characteristics cannot be fully presented through a single perspective. The traditional three-dimensional cadastral mapping technology lacks the ability to fully display the multi-dimensional information of real estate, and cannot accurately display the characteristics of complex terrain, special buildings and large-scale real estate. The management and transaction of real estate rights need to be based on accurate and reliable cadastral information support. Through the optimal perspective selection technology, the geographic information characteristics of real estate can be better presented, providing an important basis for the confirmation and transaction of rights.

[0063] Currently, when displaying three-dimensional cadastral property rights, an optimal viewing angle is often selected manually based on experience for observation. Under the optimal viewing angle determined in this way, users often cannot obtain as much visual information as possible, which affects the user experience.

[0064] To this end, an embodiment of the present invention provides a method for selecting an optimal viewing angle of a three-dimensional property body, by training a support vector machine based on a genetic algorithm through training data with corresponding scoring labels, so as to determine the optimal evaluation weights corresponding to various feature parameters that can affect the viewing angle score, and then for a target three-dimensional property body for which the optimal viewing angle needs to be selected, the feature parameter sets of the target three-dimensional property body at various viewing angles and the optimal evaluation weights corresponding to each feature parameter are calculated, and the scores of various viewing angles are determined to select the optimal viewing angle, which can ensure the scientific nature of the selection of the optimal viewing angle, improve the amount of information obtained by users at the optimal viewing angle, and ensure user experience.

[0065] According to an embodiment of the present invention, an embodiment of a method for selecting an optimal viewing angle of a three-dimensional property is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0066] In this embodiment, a method for selecting an optimal viewing angle of a three-dimensional property body is provided, which can be used for selecting an optimal viewing angle of the above-mentioned three-dimensional cadastral property body. Figure 1 FIG. 1 is a flow chart of a method for selecting an optimal viewing angle of a three-dimensional property according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0067] Step S101, obtaining a feature parameter set and a corresponding scoring label corresponding to each candidate viewing angle under multiple three-dimensional property bodies.

[0068] For the three-dimensional cadastral property body, hereinafter referred to as the three-dimensional property body, its function is to improve the display effect of the cadastral property body through the three-dimensional form. For a model in a three-dimensional state, when people watch it to understand its specific situation, they must view it from a certain perspective. There are great differences in the visual information and perception state output by different perspectives, so it is necessary to select an optimal perspective for users to view in order to improve the user experience.

[0069] The above is the application background of the embodiment of the present invention. Before determining the best viewing angle, the viewing angle of the three-dimensional property can be divided to obtain multiple candidate viewing angles, so as to select the best viewing angle from these candidate viewing angles. For the viewing angle division method, multiple fixed viewing angles can be selected based on experience or the spherical coordinate system corresponding to the three-dimensional property body can be determined, and geometric division can be performed to obtain multiple candidate viewing angles. The specific method of dividing the candidate viewing angles can be set according to the actual situation and is not limited here.

[0070] When evaluating candidate perspectives, the total score of each perspective is determined mainly based on the actual score of each feature parameter under each perspective and the weight of the influence of the feature parameter on the total score, so as to select the optimal perspective.

[0071] Therefore, in order to determine the optimal evaluation weight corresponding to each feature parameter, it is necessary to first construct a training data set and determine the optimal evaluation weight corresponding to each feature parameter through training adjustment.

[0072] Before constructing training data, it is necessary to obtain multiple three-dimensional property bodies. For each three-dimensional property body, multiple candidate perspectives are determined according to the same perspective division rule. The actual score of each perspective relative to the corresponding three-dimensional property body is determined by expert scoring. The scoring is mainly performed from the two perspectives of displayed information volume and visual perception, so as to determine the scoring label corresponding to each candidate perspective under each three-dimensional property body. Specifically, the scoring label can include five levels, referring to the five-point scale method, namely: very suitable (10 points), suitable (8 points), neutral (6 points), not very suitable (4 points), and unsuitable (2 points).

[0073] At the same time, for each candidate viewing angle of each three-dimensional property, multiple characteristic parameters under the viewing angle are determined. These characteristic parameters can be quantified according to the actual situation of the three-dimensional property. For example, the characteristic parameters may include: visibility, visible area ratio, surface area entropy, visual coherence and other quantifiable characteristic parameters of the three-dimensional property under the viewing angle. Therefore, each candidate viewing angle corresponds to a scoring label and a set of characteristic parameters, i.e., a characteristic parameter group.

[0074] Step S102: Taking the feature parameter set and the corresponding score label corresponding to each candidate perspective as a piece of training data, a training data set is constructed.

[0075] After obtaining the feature parameter sets and scoring labels of each candidate perspective corresponding to each three-dimensional property body, the feature parameter sets and scoring labels obtained in the above step S101 are sorted out, and the feature parameter sets and scoring labels corresponding to each candidate perspective are used as a training data to construct a training data set.

[0076] Step S103 , training a support vector machine based on a genetic algorithm based on the training data set, and determining the optimal evaluation weight of each feature parameter in the feature parameter set relative to the visual score.

[0077] The process of determining the optimal evaluation weights corresponding to each feature parameter for a support vector machine based on a genetic algorithm mainly involves determining the initialization population required by the genetic algorithm. Each individual in the population corresponds to a set of weights. The population is iterated multiple times through the genetic algorithm, and the optimal range of each feature parameter is determined based on the excellent individuals in the population after multiple iterations.

[0078] In the iterative process, the support of each individual in each generation of the population is determined by the support vector machine and training data, which serves as the basis for the genetic algorithm to iterate the population. After the initial range of weights is determined by the genetic algorithm, the support vector machine is used to continue multiple iterative training based on the initial range of weights, so as to finally determine the optimal evaluation weights of each feature parameter.

[0079] Step S104, obtaining a target feature parameter set corresponding to each candidate viewing angle of the target three-dimensional property, and determining a viewing angle score corresponding to each candidate viewing angle in combination with the best evaluation weight of each feature parameter relative to the visual score.

[0080] After obtaining the optimal evaluation weights corresponding to each feature parameter, the candidate perspectives of the target three-dimensional property body that needs the most optimal perspective selection can be evaluated. It is necessary to obtain the feature parameter set corresponding to each candidate perspective of the target three-dimensional property body, that is, the target feature parameter set. It should be noted that for the target three-dimensional property body, the division method of its candidate perspectives is consistent with the division method of the three-dimensional property body in the above step S101. Similarly, the determination method of each feature parameter in its feature parameter set is also consistent with the determination method of the feature parameters of the three-dimensional property body in the above step S101.

[0081] For the feature parameter set of the target three-dimensional property body, each feature parameter needs to be normalized, and the pros and cons of the same feature parameter under different viewing angles are measured by scoring, so as to determine the viewing angle score corresponding to a certain viewing angle according to the optimal evaluation weight corresponding to each feature parameter.

[0082] For example, taking the characteristic parameter of visibility as an example, assuming that there are 5 candidate viewing angles (this is just an example, the actual number is usually much), where the visibility of the target three-dimensional property body at each viewing angle is: 0.4, 0.5, 0.6, 0.7, 0.8, the visibility score at the viewing angle corresponding to 0.4 can be determined as: 0; the visibility score at the viewing angle corresponding to 0.7 can be determined as 10, and the visibility scores at each viewing angle are determined according to the proportion: 0, 2.5, 5, 7.5, 10. Here is just an exemplary normalization method, the purpose of which is to measure the pros and cons of each characteristic parameter by scoring, so as to determine the comprehensive score of the corresponding viewing angle for the pros and cons of each characteristic parameter at each viewing angle and the corresponding optimal evaluation weight.

[0083] For each viewing angle, after normalizing each feature parameter to obtain the corresponding score, it is multiplied by the corresponding optimal evaluation weight and summed up to obtain the viewing angle score corresponding to each viewing angle.

[0084] For example, suppose there are five types of feature parameters in the feature parameter set, namely A, B, C, D, and E. The corresponding optimal evaluation weights are: 0.1, 0.2, 0.2, 0.3, and 0.2.

[0085] Assume that the target 3D property has 5 viewing angles, among which the score of the normalized feature parameter corresponding to viewing angle 1 is: (8, 7, 6, 5, 4), and the viewing angle score corresponding to viewing angle 1 is: 8*0.1+7*0.2+6*0.2+5*0.3+4*0.1=5.4. Similarly, the viewing angle scores corresponding to each viewing angle are determined, which will not be repeated here.

[0086] It should be noted that the above-mentioned method of normalizing the characteristic parameters to determine the scores corresponding to each characteristic parameter is only an optional implementation method for measuring the quality of the characteristic parameters, and other methods can also be used for measurement in actual situations. For example, the quality of each characteristic parameter can be determined by using different preset intervals, which is not limited here and can be set in combination with actual conditions.

[0087] Step S105, determining the best viewing angle of the target three-dimensional property object according to the viewing angle scores corresponding to the candidate viewing angles.

[0088] After determining the view scores corresponding to the candidate view points, the view point with the highest score can be determined as the optimal view point. Alternatively, several view points with higher scores can be selected as optional view points for the user to choose, and can be set in combination with actual application scenarios.

[0089] The optimal viewing angle selection method for a three-dimensional property provided in this embodiment trains a support vector machine based on a genetic algorithm through training data with corresponding score labels, thereby determining the optimal evaluation weights corresponding to various feature parameters that can affect the viewing angle score, and then for a target three-dimensional property that needs to select the optimal viewing angle, the feature parameter set of the target three-dimensional property under each viewing angle and the optimal evaluation weights corresponding to each feature parameter are calculated, and the scores of each viewing angle are determined to select the optimal viewing angle. This can ensure the scientific nature of the selection of the optimal viewing angle, improve the amount of information obtained by users under the optimal viewing angle, and ensure user experience.

[0090] According to an embodiment of the present invention, another method embodiment for selecting the best viewing angle of a three-dimensional property body is provided, which can be used for selecting the best viewing angle of the above-mentioned three-dimensional cadastral property body. Figure 2 FIG. 4 is a flow chart of another method for selecting an optimal viewing angle of a three-dimensional property according to an embodiment of the present invention. Figure 2 As shown, the process includes the following steps:

[0091] Step S201, obtaining a feature parameter set and a corresponding scoring label corresponding to each candidate perspective under multiple three-dimensional property bodies.

[0092] Specifically, each candidate perspective is determined as follows:

[0093] Step S201-a, for each three-dimensional property body, based on the center position of the three-dimensional property body and the center position of the sphere under the preset radius, the three-dimensional property body is set in the sphere under the preset radius, and the center position of the three-dimensional property body and the center position of the sphere coincide with each other.

[0094] It can be understood that when determining the candidate viewing angles of the three-dimensional property body, the candidate viewing angles can be divided by means of a spherical coordinate system. First, the center position of the three-dimensional property body is determined, and it is used as the center of a sphere. A sphere with a preset radius is generated, and the three-dimensional property body is covered in the sphere. For example, refer to Figure 3 As shown, it is an example diagram of viewpoint deployment of a three-dimensional property according to an embodiment of the present invention. The point on the surface of the sphere in the figure can be used as a viewpoint, and the point facing the center of the sphere at the viewpoint position can be used as a viewing angle.

[0095] Furthermore, for the preset radius of the sphere corresponding to the three-dimensional property body, the preset radius is determined as follows:

[0096] For each three-dimensional property body, determining a three-dimensional bounding box of the three-dimensional property body;

[0097] The preset radius of the sphere corresponding to the three-dimensional property body is determined based on the distance between each boundary point in the three-dimensional boundary frame and the center position of the three-dimensional property body.

[0098] It can be understood that the corresponding three-dimensional bounding box is determined for the three-dimensional property body. The three-dimensional bounding box can be a cuboid or a cube, or a polygon combined with the actual shape of the three-dimensional property body, and is set according to the actual situation.

[0099] For the three-dimensional bounding box, the radius of the sphere is determined according to the distance between each boundary point and the center position of the three-dimensional property body, so that each vertex of the bounding box can be located on the spherical surface of the sphere. Specifically, the preset radius of the sphere can be determined by the following formula, where (x, y, z) represents the coordinates of the boundary points of the cadastral property body model corresponding to the three-dimensional bounding box.

[0100]

[0101] Step S201-b, constructing a spherical coordinate system of a sphere corresponding to the three-dimensional property body, and determining a plurality of grid points on the upper hemisphere of the sphere according to a preset polar angle range and a preset azimuth angle range.

[0102] After determining the sphere corresponding to the three-dimensional property right, the spherical coordinate system corresponding to the sphere can be constructed. The upper half of the sphere can be used as the division range of the candidate viewing angles, and the corresponding polar angle range is The azimuth angle range is [0,2π]. The specific polar angle range and azimuth range can be set based on the center of the sphere corresponding to the three-dimensional object and the actual situation, and there is no restriction here.

[0103] After determining the polar angle range and the azimuth angle range, geometric division can be performed within the range. For example, in a practical example, when dividing the candidate viewing angles, only the 180-degree viewing angle range above the center of the sphere is focused on, and the viewing angle range is divided into 18 equal parts. Three candidate viewing angles are selected in each equal part, for a total of 54 candidate viewing angles. Therefore, the polar angle range and the azimuth angle range can be divided into corresponding proportional parts to divide the surface of the sphere into grids, for a total of 54 grid points.

[0104] For example, the specific division method may be: divide the polar angle range and the azimuth angle range into equal intervals to obtain the number of sampling points. n θ and n δ, the index i of each polar angle interval θ The corresponding polar angle θ i , that is, theta_i, the specific calculation method is:

[0105]

[0106] The index i of each azimuth interval δ , calculate the corresponding azimuth angle δ i , the specific calculation method is:

[0107]

[0108] Get a uniformly distributed spherical coordinate system grid. For each grid point (θ i ,δ i ), converted to three-dimensional coordinates: (x, y, z) = (R*sinθ i *cosδ i ,R*sinθ*sinδ i ,R*cosδ i ).

[0109] Step S201 - c : for each grid point, determine the viewing angle of the grid point toward the center of the sphere as a candidate viewing angle.

[0110] The perspective of looking at the center of the sphere at each grid point is a candidate perspective. Taking the above example, the perspectives of the center of the carton at each of the 54 grid points are 54 candidate perspectives.

[0111] Specifically, the feature parameter set includes: visibility, visible area ratio, surface area entropy, visual connectivity, viewing comfort, visual coherence and viewing angle variability.

[0112] It can be understood that the feature parameter set for each viewing angle mainly includes the feature parameters of the above seven dimensions. In actual situations, in order to ensure the evaluation effect, more feature dimensions can be set for evaluation. There is no restriction here. Among the above seven feature parameters, visibility, visible area ratio, surface area entropy, and line of sight connectivity are mainly set from the perspective of the corresponding information amount of the three-dimensional property body under the viewing angle; viewing comfort, visual coherence and viewing angle variability are mainly set from the sensory perspective of the human eye under the viewing angle.

[0113] Furthermore, for each candidate viewing angle under each three-dimensional property body, the method for obtaining each feature parameter in the feature parameter set corresponding to the candidate viewing angle is as follows:

[0114] The visibility of the candidate viewing angle is determined according to the number of visible three-dimensional boundary points under the candidate viewing angle and the total number of boundary points of the three-dimensional property body.

[0115] For visibility, it can be expressed as (V(s,w)), and the solution formula is:

[0116]

[0117] Where N(w) is the number of three-dimensional boundary points visible from a certain viewing angle w, and N(s) is the total number of boundary points of the three-dimensional property body.

[0118] The visible area ratio of the candidate viewing angle is determined according to the projection area of ​​the candidate viewing angle relative to the three-dimensional property object and the surface area of ​​the three-dimensional property object.

[0119] The visible area ratio can be expressed as (R(s,w)), and the solution formula is:

[0120]

[0121] Where A(w) is the surface area of ​​the three-dimensional cadastral property body that can be covered by the projection of the three-dimensional cadastral property body at the viewing angle w, and A(s) is the surface area of ​​the three-dimensional cadastral property body.

[0122] According to the area of ​​each visible boundary surface under the candidate viewing angle, the surface area entropy under the candidate viewing angle is determined.

[0123] For the surface area entropy, it can be expressed as (H(m)), and the solution formula is:

[0124]

[0125] Among them A irepresents the projected area of ​​the visible boundary surface, A represents the total projected area of ​​the visible boundary surface, It reflects the visibility of the boundary surface from a certain viewpoint.

[0126] The connectivity of the sight line under the candidate perspective is determined based on the length of the connected covered sight line under the candidate perspective and the total length of the parcel body of the three-dimensional property body.

[0127] For line of sight connectivity, it can be expressed as (L(s,w)), and its solution formula is:

[0128]

[0129] Where L(w) is the sum of the lengths of connected covered sight lines, and L(s) is the total length of the plot body.

[0130] The viewing comfort of the candidate viewing angle is determined according to the angles between the candidate viewing angle and the most comfortable viewing angle of the three-dimensional object relative to the longitudinal axis of the three-dimensional object.

[0131] For viewing angle comfort, it can be expressed as (D(w)), and the solution formula is:

[0132]

[0133] where d best The most comfortable viewing angle position w best The angle between the corresponding vector and the vertical vector n of the model is d w It is the angle between a candidate viewing angle w and the model vertical direction vector n.

[0134] The visual coherence of the candidate perspective is determined based on the number of consistent features of the candidate perspective and the total number of features.

[0135] For visual coherence, it can be expressed as (C(s,w)), and its solution formula is:

[0136]

[0137] Among them C s is the number of consistent features, C w is the total feature count and view variability (D(c,w)).

[0138] The visual variability of the candidate perspective is determined based on the degree of variability between the candidate perspective and other candidate perspectives in the three-dimensional property.

[0139] The viewing angle variability can be expressed as (D(c,w)), and its solution formula is:

[0140] D=Σweight(v i , vj )×f(d i , d j )

[0141] Where f(d i ,d j ) is the actual change degree of the current perspective i relative to some other perspective j, weight is the weight function corresponding to the change degree of different perspectives, and the specific function rules can be set according to the actual situation.

[0142] For the calculation process of the above-mentioned characteristic parameters, the determination method of the boundary points, boundary lines and boundary surfaces involved can be as follows: Figure 4 , which is a schematic diagram of observing point targets, line targets, and surface targets under a three-dimensional boundary according to an embodiment of the present invention.

[0143] It should be noted that the above-mentioned method of dividing perspectives and determining characteristic parameters is also applicable to the subsequent target three-dimensional property body, and will not be repeated hereafter.

[0144] Step S202: Take the feature parameter set and the corresponding score label corresponding to each candidate perspective as a piece of training data to construct a training data set. Figure 1 Step S102 of the illustrated embodiment will not be described in detail here.

[0145] Step S203 , training a support vector machine based on a genetic algorithm based on the training data set, and determining the optimal evaluation weight of each feature parameter in the feature parameter set relative to the visual score.

[0146] Specifically, in step S203, it includes:

[0147] Step S203-1, determine the initialization population corresponding to the genetic algorithm, and iteratively optimize the initialization population, where each individual in the initialization population corresponds to a weight group.

[0148] It can be understood that the initialization population required before the genetic algorithm iteration is first determined, and the population includes multiple individuals, each of which corresponds to a set of weights, that is, a weight group. The weight group of each individual can be set based on experience or randomly, as long as the sum of the weights corresponding to each weight group is 1. The number of individuals in the initialization population can be set according to actual training requirements. In an illustrative example, it can be set to 50. Iterative optimization is performed based on the initialization population until the iteration termination condition.

[0149] Step S203-2: In each iterative optimization process, the fitness of each individual in the population is determined based on the training data set and the support vector machine, and the population is iterated through a genetic algorithm based on the fitness of each individual in the population.

[0150] Before iterative optimization of the population, the fitness of each individual is determined through the training data set and the support vector machine. The fitness can be understood as configuring the weight group corresponding to a certain individual in the support vector machine, and determining the prediction effect of the support vector machine under the weight group configuration according to the label of the training data.

[0151] By determining the fitness corresponding to each individual, the population is iteratively optimized through a genetic algorithm. For example, two individuals with higher fitness are selected from the population as the paying individuals of the next generation population, and the two individuals are crossed to obtain two new individuals; at the same time, an individual is randomly selected to mutate the weights in its weight group to obtain a mutated individual. Three individuals are selected from the current population as the replaced individuals, for example, the three individuals with the lowest fitness are selected; the two new individuals and the mutated individuals are used to replace the replaced individuals in the population to obtain a new population, and a round of fitness calculation and iterative optimization are performed.

[0152] Specifically, in the above step S203-2, in each iterative optimization process, the fitness of each individual in the population is determined based on the training data set and the support vector machine, including:

[0153] Step S203-2-1, for each individual in the population, the weight group corresponding to the individual is substituted into the support vector machine, and each training data in the training data set is predicted to obtain the prediction score corresponding to each training data.

[0154] It can be understood that the weight group corresponding to each individual is configured in the support vector machine, and a set of training data is predicted by the configured support vector machine to obtain the prediction score corresponding to each training data. During the prediction process, there is no need to update other configuration parameters of the support vector machine. The purpose of this step is to compare the prediction effects corresponding to each weight group.

[0155] Step S203-2-2, calculating the fitness of the individual according to the predicted scores and score labels corresponding to each training data.

[0156] For each individual weight group, the difference between the predicted score and the score label of each training data can be accumulated and averaged, and the average can be used as the individual fitness to characterize the effect of the weight group. The better the effect, the smaller the difference between the predicted score and the score label.

[0157] Step S203-3: when the iteration termination condition of the genetic algorithm is reached, multiple qualified weight groups are determined according to the fitness of each individual in the current population.

[0158] The termination condition of the genetic algorithm can be set to a certain number of iterations or a certain degree of fitness of individuals in the population. It can be set according to the actual situation. At this time, the fitness of each individual in the current population is determined, and a number of individuals with fitness greater than a preset value are selected as qualified individuals, or the individuals with the top fitness rankings are selected as qualified individuals. The weight group corresponding to these qualified individuals is the qualified weight group.

[0159] Step S203 - 4 , training the support vector machine based on the multiple qualified weight sets and the training data, and determining the best evaluation weight of each feature parameter in the feature parameter set relative to the visual score.

[0160] Multiple qualified weight groups are input into the support vector machine, which determines the weight range of each feature parameter according to the specific weight of each qualified weight group. Thus, training is performed in combination with the training data, and the weights are adjusted within the weight range of each feature parameter during the training process. At the same time, the constraint condition is that the sum of the weights of each feature parameter is 1, and finally a set of weight groups with the best prediction is determined, in which the corresponding weights of each feature parameter are the best evaluation weights of each feature parameter relative to the visual score.

[0161] Specifically, in the above step S203-4, determining the optimal evaluation weight of each feature parameter in the feature parameter set relative to the visual score includes:

[0162] The optimal evaluation weights of visibility, visible area ratio, surface area entropy, visual connectivity, viewing angle comfort, visual coherence and viewing angle variability in the feature parameter set relative to the visual score are determined.

[0163] It can be understood that the optimal evaluation weights obtained by the above training for each feature parameter refer to the feature parameters of seven dimensions: visibility, visible area ratio, surface area entropy, visual connectivity, viewing comfort, visual coherence and viewing angle variability.

[0164] Step S204, obtaining a target feature parameter set corresponding to each candidate viewing angle of the target three-dimensional property, and determining a viewing angle score corresponding to each candidate viewing angle in combination with the best evaluation weight of each feature parameter relative to the visual score. Figure 1 Step S104 of the illustrated embodiment will not be described in detail here.

[0165] Step S205, determining the best viewing angle of the target three-dimensional property object according to the viewing angle scores corresponding to each candidate viewing angle. Figure 1 Step S105 of the illustrated embodiment will not be described in detail here.

[0166] The optimal viewing angle selection method for a three-dimensional property provided by an embodiment of the present invention trains a support vector machine based on a genetic algorithm through training data with corresponding score labels, thereby determining the optimal evaluation weights corresponding to various feature parameters that can affect the viewing angle score, and then for a target three-dimensional property that needs to select the optimal viewing angle, the feature parameter set of the target three-dimensional property under each viewing angle and the optimal evaluation weights corresponding to each feature parameter are calculated, and the scores of each viewing angle are determined to select the optimal viewing angle. This can ensure the scientific nature of the selection of the optimal viewing angle, improve the user's information acquisition at the optimal viewing angle, and ensure user experience.

[0167] In order to facilitate understanding of the above method embodiment, an example diagram of the evaluation process of the optimal viewing angle of a three-dimensional cadastral property right body is provided according to an embodiment of the present invention, as shown in FIG. Figure 5 shown.

[0168] First, determine the data set and divide it into a training set and a prediction set, wherein the training set has corresponding scoring labels, which are used to determine the optimal weights of the feature parameters; the prediction set is the feature parameters of the three-dimensional property body at various perspectives that need to be predicted according to the determined weights.

[0169] For the training set, that is, the training data set, its format can be: X = {X1, X2, ..., X n}, where X i It is the feature parameter vector corresponding to the perspective of a three-dimensional property body, with dimension d and weight label: W i_best = {W i1 ,W i2 ,……,W i8}W i_best It is the optimal weight combination corresponding to each characteristic parameter that is expected to be determined.

[0170] First, we need to determine the initial population of the genetic algorithm, P = {P1, P2, ..., P m}, where P i represents a set of weight values ​​corresponding to the ith individual, and through SVM encoding, determines the fitness of each individual in the population. The specific calculation method is: fitness(P i )=SVM.evaluate(svm_model,X i ).

[0171] Next, the genetic algorithm is used for selection: individuals with higher fitness are selected from the population as the parents of the next generation. Specifically, the parent individuals can be determined by the roulette method. The selection probability of each individual is calculated as follows:

[0172]

[0173] Next, perform the crossover of the genetic algorithm: randomly select two parent individuals P1 and P2, randomly select two crossover points crossover_point, generate two offspring individuals C1 and C2, perform weighted crossover operation, and generate new individual solutions:

[0174] C1[i]=P1[i](i<=crossover_point), C1[i]=P2[i](i>crossover_point)

[0175] C2[i]=P2[i](i<=crossover_point), C2[i]=P1[i](i>crossover_point)

[0176] Next, perform a genetic algorithm mutation: randomly select an individual P i , randomly select the weight position mutation_position to be mutated to generate the mutated individual P mutation , perform weight mutation operation:

[0177] Pmutation[mutation_position]=P[mutation_position]+N(0,σ)

[0178] Next, the genetic algorithm is replaced. Use a replacement strategy (such as retaining elite individuals or selecting an appropriate group selection strategy) to replace a part of the individuals in the original population with the new individual solution and select the elite individual: Elite = arg max (fitness (P i )), select the remaining individuals: No_elite = P-{Elite}.

[0179] When the maximum number of iterations is reached or the fitness meets the threshold, the iteration of the genetic algorithm is terminated, and multiple groups of qualified weight combinations are output and configured in the directed support machine.

[0180] Train the SVM model: svm_model = SVM.train(X,W), and finally get the best weight combination {W best1 ,W best2 ,……,W best8}.

[0181] For the prediction set, its viewpoint quality measurement function can be set to multiply the corresponding score of each feature parameter by the corresponding weight and accumulate. Finally, the viewpoint score corresponding to each viewpoint in the prediction set is determined by combining the above-determined optimal weight combination, and the optimal viewpoint of the three-dimensional cadastral property body is determined.

[0182] In actual scenarios, the perspective can also be scored directly based on the trained SVM model to select the best perspective.

[0183] In addition, based on the above method embodiment, an implementation flow diagram of determining the optimal viewing angle of a three-dimensional cadastral property right body is provided according to an embodiment of the present invention. Figure 6 shown.

[0184] After determining the optimal weight combination corresponding to each feature parameter, it is combined with the specific feature parameter calculation method to determine the viewpoint quality evaluation standard. For the three-dimensional cadastral property model library, it can display the views of each three-dimensional cadastral property under various viewing angles in real time. At the same time, combined with the layout rules of candidate viewpoints and the calculation method of feature parameters, the feature parameters corresponding to multiple candidate viewpoints can be determined. Based on the specific viewpoint quality evaluation standard, the quality of each candidate point is calculated, the optimal viewpoint is determined, and the corresponding optimal view is output for display, so that users can understand as much three-dimensional cadastral property information as possible, and at the same time have a better visual perception.

[0185] A method for selecting an optimal viewing angle of a three-dimensional property body provided by the present invention has the following advantages: the present invention considers the influence of the three-dimensional cadastral property body itself and combines the perceptual characteristics of the human eye. The optimal viewing angle selected by this method provides better spatial perception ability, making the three-dimensional form display of the real estate more comprehensive, reducing the interference of redundant information, and providing a more realistic and intuitive visual effect, helping users to more accurately understand the spatial attributes and characteristics of the real estate, and improving the perception and understanding ability of geographic information.

[0186] In this embodiment, a device for selecting the best viewing angle of a three-dimensional property is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0187] This embodiment provides a device for selecting the best viewing angle of a three-dimensional property body, such as Figure 7 As shown, including:

[0188] The data acquisition module 401 is used to acquire the feature parameter set and the corresponding scoring label corresponding to each candidate perspective under multiple three-dimensional property bodies.

[0189] The training set construction module 402 is used to construct a training data set by taking the feature parameter set and the corresponding score label corresponding to each candidate perspective as a piece of training data.

[0190] The weight determination module 403 is used to train the support vector machine based on the genetic algorithm based on the training data set, and determine the optimal evaluation weight of each feature parameter in the feature parameter set relative to the visual score.

[0191] The viewing angle scoring module 404 is used to obtain a target feature parameter set corresponding to each candidate viewing angle of the target three-dimensional property, and determine a viewing angle score corresponding to each candidate viewing angle in combination with the best evaluation weight of each feature parameter relative to the visual score.

[0192] The viewing angle determination module 405 is used to determine the best viewing angle of the target three-dimensional property object according to the viewing angle scores corresponding to each candidate viewing angle.

[0193] In some optional implementations, the weight determination module 403, when training a support vector machine based on a genetic algorithm based on a training data set to determine the optimal evaluation weight of each feature parameter in the feature parameter set relative to the visual score, includes:

[0194] Determine the initialization population corresponding to the genetic algorithm, and iteratively optimize the initialization population, where each individual in the initialization population corresponds to a weight group;

[0195] In each iterative optimization process, the fitness of each individual in the population is determined based on the training data set and the support vector machine, and the population is iterated through the genetic algorithm based on the fitness of each individual in the population;

[0196] When the iteration termination condition of the genetic algorithm is reached, multiple qualified weight groups are determined according to the fitness of each individual in the current population;

[0197] The support vector machine is trained based on a plurality of qualified weight groups and training data to determine the optimal evaluation weight of each feature parameter in the feature parameter set relative to the visual score.

[0198] In some optional implementations, during each iterative optimization process, the weight determination module 403 determines the fitness of each individual in the population based on the training data set and the support vector machine, including:

[0199] For each individual in the population, the weight group corresponding to the individual is substituted into the support vector machine, and each training data in the training data set is predicted to obtain the prediction score corresponding to each training data;

[0200] The fitness of the individual is calculated based on the predicted scores and score labels corresponding to each training data.

[0201] In some optional implementations, when the data acquisition module 401 or the perspective scoring module 404 determines each candidate perspective, the determination method of each candidate perspective is as follows:

[0202] For each three-dimensional property body, based on the center position of the three-dimensional property body and the center position of the sphere under the preset radius, the three-dimensional property body is set within the sphere under the preset radius, and the center position of the three-dimensional property body and the center position of the sphere coincide with each other;

[0203] Constructing a spherical coordinate system of a sphere corresponding to the three-dimensional property body, and determining a plurality of grid points on the upper hemisphere of the sphere according to a preset polar angle range and a preset azimuth angle range;

[0204] For each grid point, the viewing angle of the grid point toward the sphere center is determined as a candidate viewing angle.

[0205] In some optional implementations, when the data acquisition module 401 or the viewing angle scoring module 404 determines the preset radius of the sphere, the preset radius is determined in the following manner:

[0206] For each three-dimensional property body, determining a three-dimensional bounding box of the three-dimensional property body;

[0207] The preset radius of the sphere corresponding to the three-dimensional property body is determined based on the distance between each boundary point in the three-dimensional boundary frame and the center position of the three-dimensional property body.

[0208] In some optional embodiments, the feature parameter set includes: visibility, visible area ratio, surface area entropy, visual connectivity, viewing angle comfort, visual coherence, and viewing angle variability;

[0209] The weight determination module 403, when determining the optimal evaluation weight of each feature parameter in the feature parameter set relative to the visual score, includes:

[0210] The optimal evaluation weights of visibility, visible area ratio, surface area entropy, visual connectivity, viewing angle comfort, visual coherence and viewing angle variability in the feature parameter set relative to the visual score are determined.

[0211] In some optional implementations, for each candidate viewing angle under each three-dimensional property body, the method for obtaining each feature parameter in the feature parameter set corresponding to the candidate viewing angle is as follows:

[0212] Determine the visibility of the candidate perspective according to the number of visible three-dimensional boundary points under the candidate perspective and the total number of boundary points of the three-dimensional property body;

[0213] Determining a visible area ratio of the candidate viewing angle according to a projection area of ​​the candidate viewing angle relative to the three-dimensional property body and a surface area of ​​the three-dimensional property body;

[0214] According to the area of ​​each visible boundary surface under the candidate viewing angle, the surface area entropy under the candidate viewing angle is determined;

[0215] Determine the connectivity of the sight line under the candidate perspective according to the length of the connected covered sight line under the candidate perspective and the total length of the parcel body of the three-dimensional property body;

[0216] Determining the viewing comfort of the candidate viewing angle according to the angles between the candidate viewing angle and the most comfortable viewing angle of the three-dimensional object relative to the longitudinal axis of the three-dimensional object;

[0217] Determine the visual coherence of the candidate perspectives according to the number of consistent features of the candidate perspectives and the total number of features;

[0218] The visual variability of the candidate perspective is determined based on the degree of variability between the candidate perspective and other candidate perspectives in the three-dimensional property.

[0219] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0220] The optimal viewing angle selection device of the three-dimensional property body in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0221] The embodiment of the present invention also provides a computer device having the above Figure 7 The optimal viewing angle selection device of the three-dimensional property body shown.

[0222] See also Figure 8 , Figure 8 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 8 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 8 A processor 10 is taken as an example.

[0223] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0224] The memory 20 stores instructions executable by at least one processor 10, so that at least one processor 10 executes the method shown in the above embodiment.

[0225] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0226] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0227] The computer device also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Figure 8 The example of connecting through bus is taken in the following.

[0228] The input device 30 can receive input digital or character information, and generate key signal input related to the user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator bar, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device can be a touch screen.

[0229] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0230] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for selecting the best viewing angle of a three-dimensional property body, characterized in that: The method comprises: Obtaining feature parameter sets and corresponding scoring labels corresponding to each candidate perspective under multiple three-dimensional property bodies; The feature parameter set and the corresponding score label corresponding to each candidate perspective are used as a piece of training data to construct a training data set; Training a support vector machine based on a genetic algorithm based on the training data set to determine an optimal evaluation weight of each feature parameter in the feature parameter set relative to a visual score; Obtaining a target feature parameter set corresponding to each candidate viewing angle of the target three-dimensional property, and determining a viewing angle score corresponding to each candidate viewing angle in combination with an optimal evaluation weight of each feature parameter relative to a visual score; The optimal viewing angle of the target three-dimensional property object is determined according to the viewing angle scores corresponding to each candidate viewing angle.

2. The method according to claim 1, characterized in that The step of training a support vector machine based on a genetic algorithm based on the training data set to determine the optimal evaluation weight of each feature parameter in the feature parameter set relative to the visual score includes: Determine an initialization population corresponding to the genetic algorithm, and iteratively optimize the initialization population, wherein each individual in the initialization population corresponds to a weight group; In each iterative optimization process, the fitness of each individual in the population is determined based on the training data set and the support vector machine, and the population is iterated through a genetic algorithm based on the fitness of each individual in the population; When the iteration termination condition of the genetic algorithm is reached, multiple qualified weight groups are determined according to the fitness of each individual in the current population; The support vector machine is trained based on the multiple qualified weight groups and the training data to determine the optimal evaluation weight of each feature parameter in the feature parameter set relative to the visual score.

3. The method according to claim 2, characterized in that In each iterative optimization process, the fitness of each individual in the population is determined based on the training data set and the support vector machine, including: For each individual in the population, the weight group corresponding to the individual is substituted into the support vector machine, and each training data in the training data set is predicted to obtain the prediction score corresponding to each training data; The fitness of the individual is calculated according to the predicted scores and score labels corresponding to the respective training data.

4. The method according to claim 1, characterized in that The method for determining each candidate perspective is as follows: For each three-dimensional property body, based on the center position of the three-dimensional property body and the center position of the sphere under the preset radius, the three-dimensional property body is set in the sphere under the preset radius, and the center position of the three-dimensional property body and the center position of the sphere coincide with each other; Constructing a spherical coordinate system of a sphere corresponding to the three-dimensional property body, and determining a plurality of grid points on the upper hemisphere of the sphere according to a preset polar angle range and a preset azimuth angle range; For each grid point, the viewing angle of the grid point toward the sphere center is determined as a candidate viewing angle.

5. The method according to claim 4, characterized in that The preset radius is determined as follows: For each three-dimensional property volume, determining a three-dimensional bounding box of the three-dimensional property volume; The preset radius of the sphere corresponding to the three-dimensional property body is determined based on the distance between each boundary point in the three-dimensional boundary frame and the center position of the three-dimensional property body.

6. The method according to any one of claims 1 to 5, characterized in that: The characteristic parameter set includes: visibility, visible area ratio, surface area entropy, visual connectivity, viewing angle comfort, visual coherence and viewing angle variability; Determining the optimal evaluation weight of each feature parameter in the feature parameter set relative to the visual score comprises: The optimal evaluation weights of visibility, visible area ratio, surface area entropy, visual connectivity, viewing angle comfort, visual coherence and viewing angle variability in the feature parameter set relative to the visual score are determined.

7. The method according to claim 6, characterized in that For each candidate viewing angle under each three-dimensional property body, the method for obtaining each feature parameter in the feature parameter set corresponding to the candidate viewing angle is as follows: Determining the visibility of the candidate viewing angle according to the number of visible three-dimensional boundary points under the candidate viewing angle and the total number of boundary points of the three-dimensional property body; Determining a visible area ratio of the candidate viewing angle according to a projection area of ​​the candidate viewing angle relative to the three-dimensional property body and a surface area of ​​the three-dimensional property body; Determining the surface area entropy at the candidate viewing angle according to the area of ​​each visible boundary surface at the candidate viewing angle; Determine the connectivity of the sight line under the candidate viewing angle according to the length of the connected covered sight line under the candidate viewing angle and the total length of the parcel body of the three-dimensional property body; Determining the viewing comfort of the candidate viewing angle according to the angles between the candidate viewing angle and the most comfortable viewing angle of the three-dimensional object relative to the longitudinal axis of the three-dimensional object; Determining visual coherence of the candidate perspectives according to the number of consistent features of the candidate perspectives and the total number of features; The visual variability of the candidate viewing angle is determined based on the degree of variability between the candidate viewing angle and other candidate viewing angles in the three-dimensional property.

8. A device for selecting the best viewing angle of a three-dimensional property body, characterized in that: The device comprises: A data acquisition module, used to obtain a feature parameter set and a corresponding scoring label corresponding to each candidate perspective under multiple three-dimensional property bodies; A training set construction module is used to construct a training data set by taking the feature parameter set and the corresponding score label corresponding to each candidate perspective as a piece of training data; A weight determination module, used for training a support vector machine based on a genetic algorithm based on the training data set, and determining an optimal evaluation weight of each feature parameter in the feature parameter set relative to a visual score; A viewing angle scoring module is used to obtain a target feature parameter set corresponding to each candidate viewing angle of the target three-dimensional property object, and determine a viewing angle score corresponding to each candidate viewing angle in combination with the best evaluation weight of each feature parameter relative to the visual score; The viewing angle determination module is used to determine the best viewing angle of the target three-dimensional property object according to the viewing angle scores corresponding to each candidate viewing angle.

9. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the optimal viewing angle selection method for a three-dimensional property body according to any one of claims 1 to 7 by executing the computer instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the method for selecting the best viewing angle for a three-dimensional property according to any one of claims 1 to 7.

Images