Scene model construction system and scene model constructing method

This application claims priority to China Application Serial Number 201910664639.2 filed Jul. 11, 2019, which is herein incorporated by reference in its entirety.

The present disclosure relates to a scene model construction system and scene model constructing method, in particular, a technique for classifying multiple scene materials and establishing a three-dimensional model.

Recently, there are multiple methods for constructing a three-dimensional scene model, but these methods require operation on a large amount of data, and the requirements for software and hardware are extremely high, so it is difficult to popularize the general user. In addition, when collecting a large amount of scene materials (e.g., images) to construct a three-dimensional model, there may be errors between each scene material due to different coordinate parameters, so it is difficult to combine. Therefore, there is a need for a construction technique that is more efficient and improves the accuracy of combination.

One aspect of the present disclosure is a scene model constructing method, including the following steps: Classifying multiple scene materials into multiple position groups according to multiple position parameters in the multiple scene materials. The scene materials include in a first position group of the multiple position groups as multiple first group data, and the scene materials included in a second position group of the multiple position groups as multiple second group data. Classifying the multiple first group data into multiple first similar image sub-groups as multiple first sub-group data according to scene similarities between the multiple first group data. Classifying the multiple second group data into multiple second similar image sub-groups as multiple second sub-group data according to scene similarities between the multiple second group data. Establishing multiple first similar image sub-models according to the multiple first similar image sub-groups, and establishing multiple second similar image sub-models according to the multiple second similar image sub-groups. Combining the multiple first similar image sub-models to a first position model, and combining the multiple second similar image sub-models to a second position model. Combining the first position model and the second position model to a scene model.

Another aspect of the present disclosure is a scene model construction system, including multiple terminal devices and a processor. The terminal devices are configured to transmit multiple frame data. The processor is configured to receive the multiple frame data to obtain multiple scene materials. The processor is configured to classify the multiple scene materials into multiple position groups according to multiple position parameters in the multiple scene materials, then classify the multiple position groups into multiple similar image sub-groups according to scene similarities between the multiple scene materials. The processor is configured to establish multiple similar image sub-models according to the multiple similar image sub-groups, then combine the multiple first similar image sub-models to multiple position models, and combine the multiple position models to a scene model.

Accordingly, since the present disclosure utilizes “two stages” to sequentially classify the scene materials (e.g., position, similarity), the operation time can be effectively reduced. At the same time, by sequentially combining the three-dimensional model through “two stages”, it is also possible to reduce the error that may occur when the model is combined.

It is to be understood that both the foregoing general description and the following detailed description are by examples, and are intended to provide further explanation of the disclosure as claimed.

The present disclosure can be more fully understood by reading the following detailed description of the embodiment, with reference made to the accompanying drawings as follows:

FIG. 1 is a schematic diagram of the construction system in some embodiments of the present disclosure.

FIG. 2 is a schematic diagram of scene materials in some embodiments of the present disclosure.

FIG. 3 is a schematic diagram of the constructing method in some embodiments of the present disclosure.

FIG. 4 is a flow chart of the constructing method in some embodiments of the present disclosure.

FIG. 5 is a schematic diagram of the feature vectors in some embodiments of the present disclosure.

FIG. 6 is a schematic diagram of the vocabulary tree in some embodiments of the present disclosure.

FIG. 7 is a relative relationship diagram of the vocabulary tree in some embodiments of the present disclosure.

FIG. 8A-8B are point cloud schematic diagrams of the position model in some embodiments of the present disclosure.

FIG. 9A-9C are point cloud schematic diagrams of the similar image sub-model in some embodiments of the present disclosure.

For the embodiment below is described in detail with the accompanying drawings, embodiments are not provided to limit the scope of the present disclosure. Moreover, the operation of the described structure is not for limiting the order of implementation. Any device with equivalent functions that is produced from a structure formed by a recombination of elements is all covered by the scope of the present disclosure. Drawings are for the purpose of illustration only, and not plotted in accordance with the original size.

It will be understood that when an element is referred to as being “connected to” or “coupled to”, it can be directly connected or coupled to the other element or intervening elements may be present. In contrast, when an element to another element is referred to as being “directly connected” or “directly coupled,” there are no intervening elements present. As used herein, the term “and/or” includes an associated listed items or any and all combinations of more.

Referring to FIG. 1 and FIG. 2 , FIG. 1 is a schematic diagram of a construction system 100 in some embodiments of the present disclosure. FIG. 2 is a schematic diagram of the constructing method in some embodiments of the present disclosure. The construction system 100 is configured to receive multiple scene materials in order to establish a corresponding scene model.

In some embodiments, the construction system 100 includes a server 110 and multiple terminal devices 120 . The server 110 connects to the terminal devices 120 through Internet N. Referring to FIG. 3 , the terminal devices 120 may be implemented in or implemented by a smart phone, a network device, a personal computer or devices with network connection capability. The terminal devices 120 is configured to upload multiple frame data Din 1 , Din 2 and Din 3 to Internet N, and transmit frame data Din 1 , Din 2 and Din 3 to the server 110 through Internet. The frame data Din 1 , Din 2 and Din 3 can be different images, or be a film including multiple images.

The server 110 includes a processor 111 configured to receive frame data Din 1 , Din 2 and Din 3 from the terminal device 120 , and obtain multiple scene materials Din from the frame data Din 1 , Din 2 and Din 3 . Referring to FIG. 2 , in some embodiments, after the processor 111 of the server 110 receives the frame data Din 1 , Din 2 and Din 3 , the processor 111 compares two adjacent frame data in the frame data Din 1 , Din 2 and Din 3 in order to generate multiple change values (e.g., the change value between Din 1 and Din 2 , or the change value between Din 2 and Din 3 ). The processor 111 of the server 110 generates scene materials according to the change values. For example, the difference between the frame data Din 1 , Din 2 extremely small, the difference between of the frame data Din 2 , Din 3 is obvious. When the processor 111 of the server 110 determines that the change value (difference) is larger than a threshold value, the processor 111 will reserve the frame data Din 2 and Din 3 as scene materials Din. Because the frame data Din 1 is similar with the frame data Din 2 , so the frame data Din 1 can be removed. Accordingly, the present disclosure can reduce the operation load of the server 110 during constructing model.

Each of the scene materials Din includes a position parameter. The position parameter can be a coordinate according to Global Positioning System, or can be a latitude and longitude. The position parameter is configured to represent the location corresponding to the display content of the scene materials Din. The processor 111 of the server 110 classifies the scene materials Din into multiple position groups G 1 , G 2 according to position parameters (this step is also known as Location Clustering), then classifies position groups G 1 , G 2 into multiple similar image sub-groups G 1 A, G 1 B, G 2 A and G 2 B according to the scene similarities between scene materials Din. The processor 111 of the server 110 is further configured to establish multiple similar image sub-models M 1 A, M 1 B, M 2 A and M 2 B according to the similar image sub-groups G 1 A, G 1 B, G 2 A and G 2 B. Then, combining the similar image sub-models M 1 A, M 1 B, M 2 A and M 2 B to multiple position models ML 1 , ML 2 . Finally, the processor 111 of the server 110 can combine the position models ML 1 , ML 2 to a scene model Mout.

As shown in FIG. 2 , the server 110 first performs a first grouping of the scene materials Din according to the position parameters. Then, performing a second grouping according to the similarities between the clustered scene materials Din. Through the above technical of “two stages”, the scene materials Din may be classified into multiple image sub-groups in sequence, data may be integrated more accurately and efficiently, and operation time may be reduced. At the same time, during combining the three-dimensional model, it similarly using the technical of “two stages”. It means, the small similar image sub-models are combined first, then the medium position models is combined. Finally, the scene model Mout is combined. Accordingly, the constructing method of the scene model Mout can more efficiently, and the operation load can be reduced.

In addition, the above the constructing method not only improves the operation, but also user can shoot any scene by terminal device 120 as a scene material Din and upload to the server 110 , so that a normal user can construct scene model Mout by the server 110 through Internet N.

To illustrate the implementation details of the present disclosure, as shown in FIG. 2 , In some embodiments, the scene materials Din included in the first position group GL 1 of the position groups as multiple first group data D 1 . The scene materials Din included in the second position group GL 2 of the position groups as multiple second group data D 2 . In addition, the similar image sub-groups include multiple first similar image sub-groups G 1 A, G 1 B and multiple second similar image sub-groups G 2 A, G 2 B. The first similar image sub-group G 1 A, G 1 B are configured to establish the first similar image sub-models M 1 A, M 1 B. The second similar image sub-group G 2 A, G 2 B are configured to establish the second similar image sub-models M 2 A, M 2 B. The position models include a first position model ML 1 and a second position model ML 2 .

As mentioned above, the processor 111 of the server 110 classifies the first group data D 1 into the first similar image sub-group G 1 A, G 2 B as first sub-group data D 1 A, D 1 B according to multiple scene similarities between first group data D 1 . Similarly, the processor 111 is further configured to classifies the second group data D 2 into the second similar image sub-group G 2 A, G 2 B as the second sub-group data D 2 A, D 2 B according to the scene similarities between the second position group data D 2 . The first sub-group data D 1 A, D 1 B and the second sub-group data D 2 A, D 2 B are same as the scene material Din, all of them are frame data, but are classified in different groups. In order to clearly illustrate each step of the present disclosure, it is referred to herein by different terms.

Referring to FIG. 4 , FIG. 4 is a flow chart of the constructing method in some embodiments of the present disclosure. The steps of the constructing method of the present disclosure are described below. In the step S 401 , the server 110 receives multiple scene materials Din transmitted from the terminal devices 120 . In some embodiments, the scene materials Din are transmitted by the different terminal device 120 to the server 110 at different time.

In step S 402 , the server 110 classifies scene materials Din into multiple position groups according to the position parameters of the scene material Din. As shown in FIG. 1 , in some embodiments, the server 110 includes the processor 111 , a storage unit 112 and a transmission unit 113 . The storage unit 112 is configured to store multiple region data. each of the region data is configured to represent a position parameter of a default region (e.g., “block A”, or GPS coordinates). Each of the region data corresponds to a position group. After the transmission unit 113 transmits the received scene materials Din to the processor 111 , the processor 111 compares the position parameters of the scene materials Din with the position parameters of the region data in order to classify the scene materials Din into the corresponding position groups. As shown in the embodiments of FIG. 3 , the position groups include a first position group GL 1 and a second position group GL 2 , but the number of the position groups can be modified in some embodiments. As mentioned above, the scene materials Din in the first position group GL 1 are taken as first group data D 1 . The scene materials Din in the second position group GL 2 are taken as second group data D 2 .

In the step S 403 , the processor 111 calculates the scene similarities between multiple first group data D 1 . For example, if both of the screens of two first group data D 1 are configured to represent “school”, and there are 70% portion of the screen is the same, then the scene similarity is 70%. The calculation of the scene similarities will be explained in the following paragraphs.

In the step S 404 , the processor 111 classifies first group data D 1 into the first similar image sub-groups G 1 A, G 1 B according to scene similarities between first group data D 1 , and as the first sub-group data D 1 A, D 1 B. For example, the first similar image sub-group G 1 A represents “front door of school”, the first similar image sub-group G 1 B represents “fence of school”. In some embodiments, the processor 111 classifies the first group data D 1 , whose scene similarities is larger than the threshold value (e.g., 70%) into the same similar image sub-group. Similarly, the processor 111 is further classifies the second group data D 2 into the second similar image sub-groups G 2 A, G 2 B according to scene similarities of the second group data D 2 , and as the second sub-group data D 2 A, D 2 B.

In the step S 405 , after classifying the first similar image sub-groups G 1 A, G 1 B and the second similar image sub-groups G 2 A, G 2 B, the processor 111 establishes first similar image sub-models M 1 A, M 1 B according to the first similar image sub-groups G 1 A, G 1 B, and establishes second similar image sub-models M 2 A, M 2 B according to second similar image sub-groups G 2 A, G 2 B.

In the step S 406 , the processor 111 combines the first similar image sub-models M 1 A, M 1 B to a first position model ML 1 , and combines second similar image sub-models M 2 A, M 2 B to a second position model ML 2 . Finally, in the step S 407 , the processor 111 combines the first position model ML 1 and the second position model ML 2 to a scene model Mout.

In some embodiments, the server 110 stores the generated scene model Mout to the storage unit 112 , so as to provide internet service (e.g., construct map image or virtual image engine). The scene model Mout may be a three-dimensional map information, when the terminal device 120 transmits a location request message to the server 110 , the server 110 compares a current frame data in the location request message (e.g., an image captured by the terminal device 12 ) with the scene model Mout so as to obtain a spatial coordinate information. For example, determining the current frame data corresponds to the position groups G 1 or the position groups G 2 , then obtaining the position parameter according to the compared scene material. The server 110 transmits coordinates information to the terminal device 120 . Accordingly, user can quickly and accurately use the scene model Mout in the server 110 to determine the current location. In addition, the terminal device 120 can periodically transmit a location request message to the server 110 , so that the server 110 can track the movement track of the terminal device 120 to implement the location tracking service.

In other embodiments, when the terminal device 120 transmits a browse request message (e.g., request to browse online maps, or request to launch online virtual image engine) to the server 110 , the processor 111 may transmit a three-dimensional image data generated by the scene model Mout to the terminal device 120 according to the browse request message in order to provide the internet service.

In some embodiments, the processor 111 uses an optical flow method to calculate the change value between two adjacent frame data of multiple frame data Din 1 , Din 2 and Din 3 (e.g., prune image frames). The processor 111 selects the frame data Din 1 , Din 2 , Din 3 , whose change value is larger than the threshold value, as scene material Din (or called “Key frame”). Similarly, the processor 111 can also use optical flow method to calculate the scene similarities between multiple first group data D 1 or between multiple second group data D 2 .

In addition, the processor 111 can use feature vectors to calculate the scene similarities. Take the first group data D 1 of the first position group GL 1 as an example, the processor 111 respectively obtain feature points in each of the first group data D 1 . Those feature points correspond to multiple feature vectors. Referring to FIG. 5 , the processor 111 may respectively obtain multiple portion areas as feature points from first group data D 1 through feature information algorithm (e.g., SURF, SIFT, ORB), in order to generate feature points data D 1 X. Then, analyzing each of the feature points to feature vectors (e.g., a collection containing values of 64 or 128 dimensions) so as to obtain the feature vector data D 1 Y. Feature vector can be considered a key feature of each first group data D 1 , so the processor 111 may obtain scene similarities between multiple first group data by comparing feature points or feature vectors of the first group data.

The following describes the method to combine the similar image sub-models into the position models. Take the first similar image sub-model M 1 A, M 1 B as an example, the processor 111 determines the first similar image sub-groups G 1 A, G 1 B corresponding to the first similar image sub-models M 1 A, M 1 B. Then, finding similar parts from the first similar image sub-groups G 1 A, G 1 B, and as connection points to combine the first similar image sub-models M 1 A, M 1 B. Specifically, in some embodiments, during the processor 111 combines the first similar image sub-model M 1 A, M 1 B, the processor 111 first compares the corresponding first sub-group data D 1 A, D 1 B to determine multiple feature similarities between the first sub-group data D 1 A and the first sub-group data D 1 B. The first sub-group data D 1 A, D 1 B, whose feature similarity is larger than a default value, is taken as a connection portion between the first sub-group data D 1 A and the first sub-group data D 1 B (e.g., the first sub-group data D 1 A and the first sub-group data D 1 B has the same image “fence of school”). The processor 111 sets the first sub-group data D 1 A, D 1 B, whose feature similarity is larger than the default value, as “two-dimensional connected component”. Through two-dimensional connected component, the first similar image sub-models M 1 A, M 1 B can be combined to the first position model ML 1 . Similarly, the processor 111 can combine the second similar sub-models M 2 A, M 2 B to the second position model ML 2 in the same way.

Referring to FIG. 3 , in some embodiments, after the processor 111 calculates the scene similarities between the first group data D 1 , the processor 111 further arranges the first group data D 1 , according to scene similarities between first group data D 1 to establish a vocabulary tree according to all of the classified first sub-group data D 1 A of the first similar image sub-group G 1 A (i.e., every similar image sub-group can establish a vocabulary tree). As shown in FIG. 6 , FIG. 6 is a schematic diagram of vocabulary tree 51 in some embodiments of the present disclosure. In the vocabulary tree 51 , the more similar the first sub-group data D 1 will be arranged the closer, and each first sub-group data D 1 A is used as a two-dimensional structure node 511 - 513 in the vocabulary tree 51 . For example, the two-dimensional structure nodes 511 , 512 are closer, which means that their corresponding first sub-group data D 1 has higher similarity. Similarly, all of the first sub-group data D 1 B classified into first similar image sub-group G 1 B can be established as another vocabulary tree. The vocabulary tree is used to index each of the first sub-group data D 1 A, D 1 B in the first similar image sub-groups G 1 A, G 1 B.

As mention above, the processor 111 can use the vocabulary tree to find connection points between the first similar image sub-group G 1 A and the first similar image sub-group G 1 B to combine the first similar sub-models M 1 A, M 1 B. Referring to FIG. 7 , FIG. 7 is a schematic diagram of vocabulary tree 61 , 62 in some embodiments of the present disclosure. The vocabulary tree 61 corresponds to the first similar image sub-group G 1 A, and the vocabulary tree 62 corresponds to the first similar image sub-group G 1 B. The vocabulary tree 61 includes multiple two-dimensional structure nodes 61 A˜ 61 C and 611 ˜ 615 . The vocabulary tree 62 also includes multiple two-dimensional structure nodes 62 A˜ 62 C and 621 ˜ 624 . As mentioned above, each two-dimensional structure nodes corresponds to the first sub-group data D 1 A, D 1 B shown in FIG. 3 . When the first similar image sub-model M 1 A, M 1 B is established, the processor 111 compares all of the two-dimensional structure nodes of different vocabulary trees 61 , 62 to obtain feature similarities between of the two-dimensional structure nodes. For example, the processor 111 determines feature similarities between all of the two-dimensional structure nodes 611 in the vocabulary tree 61 , all of two-dimensional structure nodes 62 A˜ 62 C and 621 ˜ 624 of the vocabulary tree 62 . If the feature similarities is greater than the default value (e.g., 80%), the two-dimensional structure nodes can be considered the same and can be set to “the two-dimensional connected components”. As shown in FIG. 6 , Feature similarities between two-dimensional structure nodes 614 , 622 is larger than the default value, the processor 111 may set two-dimensional structure nodes 614 , 622 as the two-dimensional connected components for combining the first similar sub-model M 1 A, M 1 B.

As shown in FIG. 7 , After the processor 111 compares all the two-dimensional connected components, if the processor 111 finds that the similarity between the two-dimensional structure nodes 614 , 622 is the highest, and the similarity is greater than the default value “80%”, the processor 111 will set the two-dimensional structure nodes 614 , 622 to the two-dimensional connected components. After finding the two-dimensional connected components in the vocabulary tree 61 , 62 , it can confirm the same three-dimensional block between the first similar image sub-models M 1 A, M 2 A (e.g., a specific coordinate position) according to the two-dimensional connection points, so that first similar image sub-models M 1 A, M 2 A can be combined into the first position model ML 1 .

In the foregoing embodiment, the processor 111 calculates the feature similarities by comparing the all of two-dimensional structure nodes in the vocabulary trees 61 , 62 . In some embodiments, the processor 111 may only compare the feature similarities between the two-dimensional structure nodes at a last leaf node level of the vocabulary tree 61 , 62 (or called “leaf”). That is, the processor 111 does not need to compare the similarity between the two-dimensional structure nodes 61 A- 61 C and 62 A- 62 C and other two-dimensional structure nodes. Since two-dimensional structure nodes at a last leaf node level of vocabulary trees 61 , 62 represent an edge level of the first similar image sub-model M 1 A, this method will reduce the amount of operation of the processor 111 .

In some embodiments, first similar image sub-models M 1 A, M 1 B, second similar image sub-models M 2 A, M 2 B, the first position model ML 1 and the second position model ML 2 are three-dimensional models composed of multiple coordinate points, also known as “point cloud”. The processor 111 can establish a point cloud model through the Structure from Motion algorithm according to sub-group data D 1 A, D 1 B, D 2 A, D 2 B of the similar image sub-groups G 1 A, G 1 B, G 2 A, G 2 B. That is, similar image sub-models M 1 A, M 1 B, M 2 A, M 2 B. In the point cloud model, each coordinate point can be regarded as a three-dimensional structure node. Referring to FIG. 8A and FIG. 8B , FIG. 8A and FIG. 8B are schematic diagrams of “point cloud” of the first position model ML 1 , the second position model ML 2 . The first position model ML 1 includes multiple first three-dimensional structure nodes 71 - 73 . The second position model ML 2 includes multiple second three-dimensional structure nodes 74 - 76 .

In some embodiments, during the processor combining the first position model ML 1 and the second position model ML 2 , the processor first obtain multiple first three-dimensional structure nodes 71 ˜ 73 and multiple second three-dimensional structure nodes 74 ˜ 76 in the first position model ML 1 and the second position model ML 2 . Then, comparing first three-dimensional structure nodes 71 ˜ 73 with the three-dimensional structure nodes 74 ˜ 76 to obtain the first spatial similarities between the first three-dimensional structure nodes 71 ˜ 73 and the second three-dimensional structure nodes 74 ˜ 76 . For example, the processor 111 determines the spatial similarity between first three-dimensional structure nodes 71 and second three-dimensional structure nodes 73 . If the first spatial similarity is larger than the default value (e.g., 85%), setting the first three-dimensional structure node 71 and the second three-dimensional structure node 73 as a three-dimensional connection point. After obtain the three-dimensional connection points between the first position model ML 1 and the second position model ML 2 , the processor 111 combines the first position model ML and the second position model ML 2 to a scene model Mout.

In some embodiments, during the processor 111 combines the first position model ML 1 and the second position model ML 2 , the processor can only select three-dimensional structure nodes at an edge level of first position model ML 1 and the second position model ML 2 to comparison because the combine region between of the first position model ML 1 and the second position model ML 2 is usually the peripheral area. That is, the first three-dimensional structure nodes compared by the processor are located at an edge level of the first position model ML 1 , and the second three-dimensional structure nodes are located at an edge level of the second position model ML 2 .

In addition, when the processor combines the first position model ML 1 and the second position model ML 2 according to three-dimensional connection points, since all of the distance, proportion and direction between the first position model ML 1 and the second position model ML 2 may be differences, even if the first position model ML 1 , second position model ML 2 is combined with the corresponding three-dimensional connection points, the combined three-dimensional model may possibly have errors. Therefore, in some embodiments, the processor 111 can modify the scene model Mout according to similar image sub-models.

Referring to FIGS. 9A-9C , FIGS. 9A-9C are schematic diagrams of first similar image sub-model M 1 A and second similar image sub-model M 2 A. The first similar image sub-model M 1 A includes multiple three-dimensional structure nodes 81 - 83 , the second similar image sub-model M 2 A also includes multiple three-dimensional structure nodes 84 - 86 .

During the processor 111 combines first similar image sub-models M 1 A, M 1 B to the first position model ML 1 , and combines second similar image sub-models M 2 A, M 2 B to the second position model ML 2 , and combines the first position model ML 1 and the second position model ML 2 according to three-dimensional connection points, the processor 111 is further configured to obtain substantially the same multiple three-dimensional structure nodes of first similar image sub-models M 1 A, M 1 B and second similar image sub-models M 2 A, M 2 B. the determination method of “substantially the same multiple three-dimensional structure nodes” as described above, It can choose three-dimensional structure nodes with higher spatial similarities between each other. For example, the three-dimensional structure nodes 81 - 83 and 84 - 86 shown in FIGS. 9A and 9B .

As mentioned above, the processor 111 modifies a spatial configuration of the first position model ML 1 and the second position model ML 2 (e.g., modify the coordinate direction, or change the proportion of spatial structure) according to the obtained same three-dimensional structure nodes 81 ˜ 83 and 84 ˜ 86 , so that the first position model ML 1 and the second position model ML 2 correspond to each other according to the same multiple three-dimensional structure nodes. As shown in FIG. 9C , if the three-dimensional structure nodes 81 are substantially the same as the three-dimensional structure nodes 84 , the processor 111 can calculate the connection method (e.g., rotate the first position model ML 1 by an adjustment angle R) through an algorithm (e.g., the Iterative Closest Point algorithm), so that the corresponding three-dimensional structure nodes 81 , 84 can be coincide. Accordingly, the first position model ML 1 and the second position model ML 2 can be accurately combined to the scene model Mout.

Alternatively stated, when combining according to three-dimensional connection points between the first position model ML 1 and the second position model ML 2 , there may be distance, direction error between the first position model ML 1 and the second position model ML 2 due to different coordinate parameters. Accordingly, by further comparing similar image sub-models, and modifying according to the substantially same three-dimensional structure nodes between similar image sub-models, the error will be corrected to correctly generate the scene model Mout.

The elements, method steps, or technical features in the foregoing embodiments may be combined with each other, and are not limited to the order of the specification description or the order of the drawings in the present disclosure.

It will be apparent to those skilled in the art that various modifications and variations can be made to the structure of the present disclosure without departing from the scope or spirit of the present disclosure. In view of the foregoing, it is intended that the present disclosure cover modifications and variations of this present disclosure provided they fall within the scope of the following claims.