阅读说明：本技术 一种点群数据特定目标数据提取的自动建模方法及系统 (Automatic modeling method and system for extracting specific target data of point group data ) 是由 关键 邹洋 于 2021-06-24 设计创作，主要内容包括：本发明涉及点群数据处理技术领域,提供一种点群数据特定目标数据提取的自动建模方法及系统,所述方法包括：步骤100,从点群数据获取设备得到的点群数据,对点群数据中所有数据点进行RGB空间到HSL空间的转换；步骤200,依据数据点的色相角属性值进行色相区域分类,进行色相角区间分类；根据色相角区间分类的情况与目标物本身颜色特征,在色相角区间分类中选取目标物色相角区间范围,进而确定所有的目标点群；步骤300,在所有点群中划分目标点群总集、候选核心点群和后补点群,并生成近似目标核心点群；步骤400,对生成近似目标核心点群,进行自动建模,得到预期的目标物模型。本发明能够实现自动建模,增强建模过程的鲁棒性。(The invention relates to the technical field of point group data processing, and provides an automatic modeling method and system for extracting specific target data of point group data, wherein the method comprises the following steps: step 100, converting all data points in the point group data from an RGB space to an HSL space from the point group data obtained by the point group data acquisition equipment; step 200, classifying hue areas according to hue angle attribute values of data points, and classifying hue angle intervals; selecting a target object hue angle interval range in the hue angle interval classification according to the hue angle interval classification condition and the color characteristics of the target object, and further determining all target point groups; step 300, dividing a target point cluster total set, a candidate core point group and a post-complementing point group in all the point groups, and generating an approximate target core point group; and 400, automatically modeling the generated approximate target core point group to obtain an expected target object model. The invention can realize automatic modeling and enhance the robustness of the modeling process.)

1. An automatic modeling method for extracting specific target data of point group data is characterized by comprising the following steps:

step 100, converting all data points in the point group data from an RGB space to an HSL space from the point group data obtained by the point group data acquisition equipment;

step 200, classifying hue areas according to hue angle attribute values of data points, and classifying hue angle intervals; selecting a target object hue angle interval range in the hue angle interval classification according to the hue angle interval classification condition and the color characteristics of the target object, and further determining all target point groups;

step 300, dividing the target point cluster total set M1 and the candidate core point cluster M in all the point clusters_KAnd the postsynaptic M group_reAnd generating an approximate target core point group M_{app_F}The method comprises steps 301 to 303:

step 301, dividing the target point cluster total M1 and the candidate core point cluster M in all the point clusters_KAnd the postsynaptic M group_re；

Step 302, for the candidate core point group M_KExpanding to obtain a space core point group M_app: for candidate core point group M_KAll points in the interior are classified as a space core point group M_app(ii) a For candidate core point group M_KAny one of the core data points n_iDetermining the core data point n by using the three-dimensional Euler distance_iAnd the data point n to be determined in the target point cluster total M1_jWhen the distance is less than the core data point n_iWhen the neighborhood is not determined, the data point n to be determined is determined_jGrouped into spatial core points M_app；

Step 303, traverse the post-compensation point group M_reAll post-complement data points n in (1)_sWhen the data point n is complemented_sLess than the post-complement data point n_sTo spatial core point group M_appThe post-complement data point n is used as the neighborhood of_sFilling-in space core point group M_appGenerating an approximate target core point group M_{app_F}；

Step 400, generating approximate target core point group M_{app_F}And automatically modeling to obtain an expected target model.

2. The method of claim 1, wherein the property values of the transformed data points comprise three-dimensional spatial coordinates and RGB color components, and the data points are denoted as n_i(x_i，y_i，z_i，r_i，g_i，b_i)；n_iIs shown asi data points; x is the number of_i，y_i，z_iRespectively representing the X-axis, Y-axis and Z-axis components of the spatial coordinates of the ith data point; r is_i，g_i，b_iRespectively representing the red, green and blue characteristic values of the RGB color gamut of the ith data point.

The converted data point attribute value comprises a space three-dimensional coordinate and an HSL component, and the data point is marked as n_i(x_i，y_i，z_i，h_i，s_i，l_i) (ii) a Wherein h is_iRepresenting the hue angle of the ith data point; s_iRepresents the saturation of the ith data point; l_iIndicating the intensity of the ith data point.

3. The method for automatic modeling of point cloud data-specific target data extraction according to claim 2, wherein in step 100, the conversion from RGB space to HSL space is performed according to the following formula;

wherein: h represents a characteristic value of a hue angle, h belongs to [0,360 ]; s represents a saturation characteristic value, and s belongs to [0,1 ]; l represents a luminance characteristic value, and l belongs to [0,1 ]; r represents a characteristic value of red, and r belongs to [0,255 ]; g represents a green characteristic value, and g belongs to [0,255 ]; b represents a blue characteristic value, b ∈ [0,255 ]; max represents the maximum value of all dot data color components (r, g, b), max ∈ [0,255 ]; min represents the minimum of all dot data color components (r, g, b), and min ∈ [0,255 ].

4. The method of claim 1, wherein in step 200, the classification of hue angle intervals is performed according to the following formula:

D＝360/B；

wherein D represents the classification number of the hue angle interval, and the value is a positive natural number; b represents an interval standard value, and the value of B is a positive natural number.

5. The method of claim 1, wherein in step 300, the target point cluster total M1 and the candidate core point cluster M are divided among all the point clusters_KAnd the postsynaptic M group_reThe process of (2) is as follows:

selecting K data points from all the target point groups as candidate core point groups M_K(ii) a Removing candidate core point group M from all target point groups_KForming a target point cluster total M1; removing candidate core point group M_KForming a post-compensation point cluster M with the point clusters outside the target point cluster total M1_re。

6. The method for automatic modeling of point cloud data-specific object data extraction as claimed in claim 1, wherein in step 400, automatic modeling is performed using Delaunay triangulation.

7. An automated modeling system for point cloud data-specific target data extraction, comprising: the system comprises an initial point cloud data module, a target point group determining module, a target object data extracting module and an automatic modeling module;

the initial point cloud data module is used for converting an RGB space to an HSL space of all data points in the point cloud data from the point cloud data obtained by the point cloud data acquisition equipment;

the target point group determining module is used for classifying hue regions according to the hue angle attribute values of the data points and classifying hue angle intervals; selecting a target object hue angle interval range in the hue angle interval classification according to the hue angle interval classification condition and the color characteristics of the target object, and further determining all target point groups;

the target object data extraction module divides a target point cluster total set M1 and a candidate core point cluster M in all the point clusters_KAnd the postsynaptic M group_reAnd generating an approximate target core point group M_{app_F}The method comprises the following steps: the point group divides submodule, data first completion submodule and data second completion submodule:

the point cluster dividing submodule divides a target point cluster total set M1 and a candidate core point cluster M in all the point clusters_KAnd the postsynaptic M group_re；

The data one-time completion submodule is used for performing M on the candidate core point group_KExpanding to obtain a space core point group M_app: for candidate core point group M_KAll points in the interior are classified as a space core point group M_app(ii) a For candidate core point group M_KAny one of the core data points n_iDetermining the core data point n by using the three-dimensional Euler distance_iAnd the data point n to be determined in the target point cluster total M1_jWhen the distance is less than the core data point n_iWhen the neighborhood is not determined, the data point n to be determined is determined_jGrouped into spatial core points M_app；

The secondary data completion submodule is a traversal post-completion point group M_reAll post-complement data points n in (1)_sWhen the data point n is complemented_sLess than the post-complement data point n_sTo spatial core point group M_appThe post-complement data point n is used as the neighborhood of_sFilling-in space core point group M_appGenerating an approximate target core point group M_{app_F}；

The automatic modeling module generates an approximate target core point group M_{app_F}And automatically modeling by adopting a Delaunay triangulation method to obtain an expected target model.

Technical Field

The invention relates to the technical field of point cloud data processing, in particular to an automatic modeling method and system for extracting specific target data of point cloud data.

Background

The three-dimensional point group measuring equipment can obtain a large amount of point data, and the point group data has the characteristics of large data volume and diversified information. Research to model based on acquired point cloud data is a major research direction in the industry.

The existing common modeling method adopts manual modeling of operators to meet expected requirements, but the manual modeling efficiency is low, manual intervention is more, and automatic modeling cannot be realized at present. How to find a technical scheme which can realize the automatic modeling from the point group data target data extraction to the target and obtain the expected effect is a problem and a difficulty which are urgently needed to be solved.

Disclosure of Invention

The invention mainly solves the technical problems that the conventional common modeling method is manual modeling of operators, the manual modeling efficiency is low, manual intervention is more, and automatic modeling cannot be realized at present, and provides an automatic modeling method and system for extracting specific target data of point group data, so as to realize automatic modeling and enhance the robustness of the modeling process.

The invention provides an automatic modeling method for extracting specific target data of point group data, which comprises the following steps:

step 100, converting all data points in the point group data from an RGB space to an HSL space from the point group data obtained by the point group data acquisition equipment;

step 200, classifying hue areas according to hue angle attribute values of data points, and classifying hue angle intervals; selecting a target object hue angle interval range in the hue angle interval classification according to the hue angle interval classification condition and the color characteristics of the target object, and further determining all target point groups;

step 300, dividing the target point cluster total set M1 and the candidate core point cluster M in all the point clusters_KAnd the postsynaptic M group_reAnd generating an approximate target core point group M_{app_F}The method comprises steps 301 to 303:

step 301, dividing the target point cluster total M1 and the candidate core point cluster M in all the point clusters_KAnd the postsynaptic M group_re；

Step 302, for the candidate core point group M_KExpanding to obtain a space core point group M_app: for candidate core point group M_KAll points in the interior are classified as a space core point group M_app(ii) a For candidate core point group M_KAny one of the core data points n_iDetermining the core data point n by using the three-dimensional Euler distance_iAnd the data point n to be determined in the target point cluster total M1_jWhen the distance is less than the core data point n_iWhen the neighborhood is not determined, the data point n to be determined is determined_jGrouped into spatial core points M_app；

Step 303, traverse the post-compensation point group M_reAll post-complement data points n in (1)_sWhen the data point n is complemented_sLess than the post-complement data point n_sTo spatial core point group M_appThe post-complement data point n is used as the neighborhood of_sFilling-in space core point group M_appGenerating an approximate target core point group M_{app_F}；

Step 400, generating approximate target core point group M_{app_F}And automatically modeling to obtain an expected target model.

Further, in step 100, the attribute values of the data points before conversion include three-dimensional spatial coordinates and RGB color components, and the data points are denoted as n_i(x_i，y_i，z_i，r_i，g_i，b_i)；n_iRepresents the ith data point; x is the number of_i，y_i，z_iRespectively representing the X-axis, Y-axis and Z-axis components of the spatial coordinates of the ith data point; r is_i，g_i，b_iRespectively representing the red, green and blue characteristic values of the RGB color gamut of the ith data point.

The converted data point attribute value comprises a space three-dimensional coordinate and an HSL component, and the data point is marked as n_i(x_i，y_i，z_i，h_i，s_i，l_i) (ii) a Wherein h is_iRepresenting the hue angle of the ith data point; s_iRepresents the saturation of the ith data point; l_iIndicating the intensity of the ith data point.

Further, in step 100, the RGB space is converted into the HSL space according to the following formula;

wherein: h represents a characteristic value of a hue angle, h belongs to [0,360 ]; s represents a saturation characteristic value, and s belongs to [0,1 ]; l represents a luminance characteristic value, and l belongs to [0,1 ]; r represents a characteristic value of red, and r belongs to [0,255 ]; g represents a green characteristic value, and g belongs to [0,255 ]; b represents a blue characteristic value, b ∈ [0,255 ]; max represents the maximum value of all dot data color components (r, g, b), max ∈ [0,255 ]; min represents the minimum of all dot data color components (r, g, b), and min ∈ [0,255 ].

Further, in step 200, the classification of hue angle intervals is performed according to the following formula:

D＝360/B；

wherein D represents the classification number of the hue angle interval, and the value is a positive natural number; b represents an interval standard value, and the value of B is a positive natural number.

Further, in step 300, the target point cluster total M1 and the candidate core point cluster M are divided among all the point clusters_KAnd the procedure for the late patch group Mre is as follows:

selecting K data points from all the target point groups as candidate core point groups M_K(ii) a Removing candidate core point group M from all target point groups_KForming a target point cluster total M1; removing candidate core point group M_KForming a post-compensation point cluster M with the point clusters outside the target point cluster total M1_re。

Further, in step 400, automatic modeling is performed using Delaunay triangulation.

Correspondingly, the invention also provides an automatic modeling system for extracting the specific target data of the point group data, which comprises the following steps: the system comprises an initial point cloud data module, a target point group determining module, a target object data extracting module and an automatic modeling module;

the initial point cloud data module is used for converting an RGB space to an HSL space of all data points in the point cloud data from the point cloud data obtained by the point cloud data acquisition equipment;

the target point group determining module is used for classifying hue regions according to the hue angle attribute values of the data points and classifying hue angle intervals; selecting a target object hue angle interval range in the hue angle interval classification according to the hue angle interval classification condition and the color characteristics of the target object, and further determining all target point groups;

the target object data extraction module divides a target point cluster total set M1 and a candidate core point cluster M in all the point clusters_KAnd the postsynaptic M group_reAnd generating an approximate target core point group M_{app_F}The method comprises the following steps: the point group divides submodule, data first completion submodule and data second completion submodule:

the point cluster dividing submodule divides a target point cluster total set M1 and a candidate core point cluster M in all the point clusters_KAnd the postsynaptic M group_re；

The data one-time completion submodule is used for performing M on the candidate core point group_KExpanding to obtain a space core point group M_app: for candidate core point group M_KAll points in the interior are classified as a space core point group M_app(ii) a For candidate core point group M_KAny one of the core data points n_iDetermining the core data point n by using the three-dimensional Euler distance_iAnd the data point n to be determined in the target point cluster total M1_jWhen the distance is less than the core data point n_iWhen the neighborhood is not determined, the data point n to be determined is determined_jGrouped into spatial core points M_app；

The secondary data completion submodule is a traversal post-completion point group M_reAll post-complement data points n in (1)_sWhen the data point n is complemented_sLess than the post-complement data point n_sTo spatial core point group M_appThe post-complement data point n is used as the neighborhood of_sFilling-in space core point group M_appGenerating an approximate target core point group M_{app_F}；

The automatic modeling moduleFor generating the approximate target core point group M_{app_F}And automatically modeling by adopting a Delaunay triangulation method to obtain an expected target model.

The automatic modeling method and the automatic modeling system for extracting the specific target data of the point group data have the advantages of definite target, strong accuracy and high efficiency. Because the point group data is sampled from the point data of the actual object, and the actual object may have color missing (such as paint falling, covering and the like), the invention adopts a secondary completion method in the modeling process, and the color feature missing data in the point group data can be re-extracted by using the method, thereby enhancing the robustness of the modeling process.

Drawings

FIG. 1 is a flow chart of an implementation of an automatic modeling method for group data specific target data extraction provided by the present invention;

FIG. 2 is a schematic illustration of a hue angle interval classification;

FIG. 3 is a schematic diagram of a cluster population of target points;

FIG. 4 is a schematic diagram of a target point cluster total, a candidate core point cluster, and a late complement point cluster;

FIG. 5 is a schematic diagram of a spatial core point group;

FIG. 6 is a schematic diagram of an approximate target core point group;

FIG. 7 is a schematic diagram of a results model;

FIG. 8 is a block diagram of a modular connectivity for an automated modeling system for crowd-data specific target data extraction provided by the present invention.

Detailed Description

In order to make the technical problems solved, technical solutions adopted and technical effects achieved by the present invention clearer, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some but not all of the relevant aspects of the present invention are shown in the drawings.

Example one

As shown in fig. 1, the automatic modeling method for extracting specific target data of point cloud data provided by the embodiment of the present invention includes the following steps:

step 100, converting all data points in the point group data from the point group data acquisition device to the HSL space from the RGB space.

In this step, point group data is obtained from the point group data acquisition device, and the number of points in the point group data is N. The attribute values of the data points before conversion include spatial three-dimensional coordinates and RGB color components, and the data points are denoted as n_i(x_i，y_i，z_i，r_i，g_i，b_i)；n_iRepresents the ith data point; x is the number of_i，y_i，z_iRespectively representing the X-axis, Y-axis and Z-axis components of the spatial coordinates of the ith data point; r is_i，g_i，b_iRespectively representing the red, green and blue characteristic values of the RGB color gamut of the ith data point.

The converted data point attribute value comprises a space three-dimensional coordinate and an HSL component, and the data point is marked as n_i(x_i，y_i，z_i，h_i，s_i，l_i) (ii) a Wherein h is_iRepresenting the hue angle of the ith data point; s_iRepresents the saturation of the ith data point; l_iIndicating the intensity of the ith data point.

In step 100, the RGB space is converted into the HSL space according to the following formula;

wherein: h represents a characteristic value of a hue angle, h belongs to [0,360 ]; s represents a saturation characteristic value, and s belongs to [0,1 ]; l represents a luminance characteristic value, and l belongs to [0,1 ]; r represents a characteristic value of red, and r belongs to [0,255 ]; g represents a green characteristic value, and g belongs to [0,255 ]; b represents a blue characteristic value, b ∈ [0,255 ]; max represents the maximum value of all dot data color components (r, g, b), max ∈ [0,255 ]; min represents the minimum of all dot data color components (r, g, b), and min ∈ [0,255 ].

Therefore, s and l do not affect the color classification, and the hue region classification is performed according to the value of the hue angle attribute value h of the dot data.

Step 200, classifying hue areas according to hue angle attribute values of data points, and classifying hue angle intervals; and selecting a target object hue angle interval range in the hue angle interval classification according to the condition of the hue angle interval classification and the color characteristics of the target object, and further determining all target point groups.

In step 200, the hue angle interval classification is performed according to the following formula, as shown in fig. 2:

D＝360/B；

wherein D represents the classification number of the hue angle interval, and the value is a positive natural number; b represents an interval standard value, and the value of B is a positive natural number. The interval standard value B can be determined and given according to the precision quality of the point group and experience.

The user selects a target object hue angle interval range from the hue angle interval classification D according to the condition of the hue angle interval classification D and the color characteristics of the target object, as shown in fig. 3. In the process, the user can observe the three-dimensional image of the selected point cluster aggregate in real time through the image display system, and the selected point is highlighted. This process has flexible variability.

Step 300, dividing the target point cluster total set M1 and the candidate core point cluster M in all the point clusters_KAnd the postsynaptic M group_reAnd generating an approximate target core point group M_{app_F}The method comprises steps 301 to 303:

step 301, dividing the target point cluster total M1 and the candidate core point cluster M in all the point clusters_KAnd the postsynaptic M group_reThe specific process is as follows:

selecting K data points from all target point groups as candidate core point groupsM_K(ii) a Removing candidate core point group M from all target point groups_KForming a target point cluster total M1; removing candidate core point group M_KForming a post-compensation point cluster M with the point clusters outside the target point cluster total M1_re。

In the step, K data points can be selected according to experience to be used as candidate core point group M_K. Thus far, as shown in FIG. 4, all the point clusters are divided into three parts, namely a target point cluster total M1 and a candidate core point cluster M_KAnd the postsynaptic M group_re。

Step 302, for the candidate core point group M_KExpanding to obtain a space core point group M_app: for candidate core point group M_KAll points in the interior are classified as a space core point group M_app(ii) a For candidate core point group M_KAny one of the core data points n_iDetermining the core data point n by using the three-dimensional Euler distance_iAnd the data point n to be determined in the target point cluster total M1_jWhen the distance is less than the core data point n_iWhen the neighborhood is not determined, the data point n to be determined is determined_jGrouped into spatial core points M_app。

Core data point n_iAnd the data point n to be determined_jDistance offset of_ijThe neighborhood centered at the selected point is denoted epsilon.

if offset_ij<epsilon determination of n_jIs n_iPoints within a point neighborhood; calculating n given the value epsilon of the neighborhood empirically_iSelecting the offset meeting in the target point cluster total M1 according to the neighborhood epsilon of the points_ij<Points of epsilon condition, classified as a spatial core point group M_appAnd marks this point as "visited". Each data point n to be determined in the target point cluster total M1_jAfter the judgment is finished, obtaining a space core point group M_appProceed to step 303. As shown in fig. 5. This step is to candidate core point group M_KThe expansion is carried out, the range of the data is limited, and the maximum degree of reduction is realizedThe operation time of the algorithm is shortened, and the modeling efficiency is improved.

Step 303, traverse the post-compensation point group M_reAll post-complement data points n in (1)_sWhen the data point n is complemented_sLess than the post-complement data point n_sTo spatial core point group M_appThe post-complement data point n is used as the neighborhood of_sFilling-in space core point group M_appGenerating an approximate target core point group M_{app_F}As in fig. 6;

in this step, for the post-compensation point group M_reAll post-complement data points n in (1)_sMiddle offset_ij<E, the points are included as graph completion repair data in the point group, namely:

e represents the population M of the post-complement data points to the spatial core points_appThe neighborhood of (a) is given by experience and practical needs; the step makes up the defect (such as paint falling of the scanned object, covering and the like) caused by the point group data_appThe data in the method is lost, so that incomplete modeling caused by insufficient data points extracted when the defects are too large can be prevented, and the modeling accuracy is improved.

Through the processing in step 300, point cloud data that meets the criteria can be extracted from a large amount of point cloud data.

Step 400, generating approximate target core point group M_{app_F}And automatically modeling to obtain an expected target model.

The method adopts a Delaunay triangulation method to carry out automatic modeling: firstly, a target core point group M is approximated_{app_F}Is constructed in a very large triangle so that M_{app_F}All points in the triangle, and secondly, inserting an approximate target core point group M one by one in a modeling system_{app_F}Every time a point is inserted, the circumcircles of existing triangles must be searched, if the inserted point is positioned inside the circumcircles of some triangles, the triangles are deleted from the triangle queue,thereby forming a polygonal cavity; finally, the insertion points and the polygonal cavities are connected to form a plurality of new triangles with the insertion points as common vertexes. Performing triangular mesh optimization according to the Delaunay triangulation empty circumcircle criterion (the circumcircle of each triangle does not contain any other points of the point set) until the standard core point group M_{app_F}All points in (a) are interpolated until the expected object envelope model is obtained, as shown in fig. 7.

In the step, model conversion from points to surfaces is carried out on the extracted point group data, so that a target object envelope model achieving the expected effect is generated.

Example two

As shown in fig. 8, the present embodiment provides an automatic modeling system for extracting point cloud data specific target data, including: the system comprises an initial point cloud data module, a target point group determining module, a target object data extracting module and an automatic modeling module;

the initial point cloud data module is used for converting an RGB space to an HSL space of all data points in the point cloud data from the point cloud data obtained by the point cloud data acquisition equipment;

the target point group determining module is used for classifying hue regions according to the hue angle attribute values of the data points and classifying hue angle intervals; selecting a target object hue angle interval range in the hue angle interval classification according to the hue angle interval classification condition and the color characteristics of the target object, and further determining all target point groups;

the target object data extraction module divides a target point cluster total set M1 and a candidate core point cluster M in all the point clusters_KAnd the postsynaptic M group_reAnd generating an approximate target core point group M_{app_F}The method comprises the following steps: the point group divides submodule, data first completion submodule and data second completion submodule:

the point cluster dividing submodule divides a target point cluster total set M1 and a candidate core point cluster M in all the point clusters_KAnd the postsynaptic M group_re；

The data one-time completion submodule is used for performing M on the candidate core point group_KExpanding to obtain a space core point group M_app: for candidate core point group M_KAll points in the interior are classified as a space core point group M_app(ii) a For candidate core point group M_KAny one of the core data points n_iDetermining the core data point n by using the three-dimensional Euler distance_iAnd the data point n to be determined in the target point cluster total M1_jWhen the distance is less than the core data point n_iWhen the neighborhood is not determined, the data point n to be determined is determined_jGrouped into spatial core points M_app；

The secondary data completion submodule is a traversal post-completion point group M_reAll post-complement data points n in (1)_sWhen the data point n is complemented_sLess than the post-complement data point n_sTo spatial core point group M_appThe post-complement data point n is used as the neighborhood of_sFilling-in space core point group M_appGenerating an approximate target core point group M_{app_F}；

The automatic modeling module generates an approximate target core point group M_{app_F}And automatically modeling by adopting a Delaunay triangulation method to obtain an expected target model.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: modifications of the technical solutions described in the embodiments or equivalent replacements of some or all technical features may be made without departing from the scope of the technical solutions of the embodiments of the present invention.