3.1. Smart Sensor Model Construction

In the system in which the camera is located, the key problem to be solved is to place the virtual objects drawn by the computer in the correct position, through the alignment and then achieve the purpose of accurate fusion, so to clarify the conversion relationship between the various coordinate systems is the key to achieve this problem. The meaning of space technology is not simply to mechanize the assembly of installations and space but to inform and media space at a deeper level. The purpose is to use the advantages of technology to create an intelligent space environment, which is fundamentally realized. Communication between people and space. The whole augmented reality system mainly involves four kinds of coordinate systems, respectively, the image plane coordinate system (image coordinate system), the camera coordinate system (video camera coordinate system), the physical coordinate system (world coordinate system), and the virtual object in the coordinate system (virtual object coordinate system). This includes the conversion between real-world and camera coordinate systems, the conversion between camera coordinate system and image plane coordinate system, and the transformation between virtual coordinate system and real coordinate system. The conversion relationship between the four coordinate systems is shown in Figure 2.

The current tension between doctors and patients is mainly due to poor communication. First, there is a huge difference in the amount of medical knowledge and medical information held by doctors and patients, which makes it sometimes difficult for patients and their families to understand the specific conditions of patients. Secondly, in traditional doctor-patient communication, patients play a passive role and lack effective interaction with doctors. The advent of mobile augmented reality technology provides excellent solutions to both problems. Using the patient’s computed tomography, magnetic resonance imaging, and other image data, the patient’s lesion model is segmented and reconstructed in three dimensions by medical image processing software, and then, the patient-specific personalized medical 3D model is presented directly to the patient or the patient’s family using the MAR system based on the mobile device side (mobile phone and tablet) [16]. Through the visual demonstration and operation of the model (pan, rotate, and zoom) and the doctor’s explanation, the patient will have a deeper understanding of the condition, and this new way of medical interaction will make the traditional doctor-patient communication more simple, clear, and direct. Most of the existing medical augmented reality systems based on mobile devices are based on the 3D registration technology of manual signs and applied to intraoperative navigation, while the 3D registration technology based on manual signs can only be applied to scenes with signs, resulting in a limited tracking area, which not only affects the scope of use of medical mobile augmented reality systems but also makes the system not convenient and stable enough. Therefore, in this paper, we choose to use the ARKit framework of SLAM technology to complete the development of a mobile augmented reality convenient display system, which can be used by medical personnel to display the lesion model for patients and their families anytime and anywhere without the limitation of markers. In addition, the system can also be applied to medical education, so that the complex structure of two-dimensional medical images on paper is more conducive to the understanding and memory of learners.

3.2. Image Virtual Reconstruction Model Design

The design of multimedia resources contains three parts: audio, video, and 3D model. Audio is an important component, and this topic selects the intelligent voice generated by reading aloud female in the processing of voice and plays the voice by scanning the text to enhance the user’s memory of the text; integrates video elements and plays the promotional video of the property by scanning the graphics of the property, which has more visual impact relative to the image; and produces a realistic virtual 3D model, so that the user can view house models and indoor roaming without leaving home [17]. Through these multimedia resources, the user can interact with the computer in real-time, which is a new model of human-computer interaction. The mobile augmented reality application consists of three major interfaces, the main interface, AR scanning interface, and indoor roaming interface; the key to interface design is icon design and interaction design; icon design should have the role of content orientation and impact while paying attention to the color and style of unity; interaction design should follow the logical relationship between the interface and the buttons. The logical relationship between interfaces is shown in Figure 3.

Image segmentation technology refers to the process of segmenting an image into some disjoint regions (segmentation of selected features showing consistency in the same region) based on features such as grayscale, color, spatial texture, and geometry and extracting the region of interest, which is the basis for fields such as image processing and computer vision. In the medical field, due to the differences in the imaging principles of medical imaging devices, the complexity of human anatomy, and the diversity of human tissue and organ shapes, the formation of images is often affected by, for example, noise, tissue motion, field offset effects, and local body effects, and thus has characteristics such as blurring and inhomogeneity, which brings great difficulties to the segmentation of medical images. To date, there is still no universal medical image segmentation technique for clinical applications, but scholars at home and abroad have reached a consensus on the general rules of image segmentation, and a considerable number of research results and methods have been produced as a result. At present, various medical image segmentation methods widely used around the world can be mainly classified into edge-based, region-based, and combined with specific theoretical image segmentation according to their segmentation characteristics. The typical ones are threshold segmentation, region growth, wavelet transform, statistics-based, and Artificial Neural Network- (ANN-) based methods. Image segmentation technology refers to the segmentation of an image into several disjoint regions based on features such as grayscale, color, spatial texture, and geometric shapes (the selected features of the segmentation show consistency in the same region), and the region of interest is extracted The process is the foundation of image processing and computer vision. Based on the theory of the above methods, two types of image segmentation have also arisen, namely, automatic segmentation and manual segmentation. Automatic intelligent segmentation is mainly done with the help of high-performance computers’ understanding of medical images and many operations to complete fully automatic image segmentation; manual segmentation requires human participation in the division and calibration in advance, and then computer operations to complete the medical image segmentation.

4.1. Smart Sensor Model Performance Analysis

One way to create a vertical three-dimensional display space model is to use the placement of mezzanine space or the hollowing out of each layer of space and other techniques to form a display space with multiple layers of “viewing platforms” in the vertical direction. The multilayered space increases the space area that can accommodate the audience and divides several clear vertical levels in the display space, which is convenient for the audience to stay. At the same time, each level is relative. Independent and visually connected, it allows the people gathered on each platform to view the exhibition from different perspectives while forming an interactive exchange between the layers, exchanging the information acquired under their respective perspectives and thus guiding the flow of visitors between the layers. For example, in the design of the Kerkrade Museum in Limburg, the architects created a half-underground, half-above-ground spherical space for the display of digital images and used the hemispherical part of the ground floor to form an inverted spherical gallery. To make full use of the space to accommodate more visitors, a circular glass platform was inserted to form a mezzanine space to accommodate the audience, and a staircase was used to form a link between the levels inside the exhibition hall, so that people can look down on the dome at different heights in the space, creating a visual experience as if they were looking back at the Earth from space, as shown in Figure 4.

The shaping of spatial differentiation is the main way for landscape installations to shape the landscape space and stimulate public participation in the experience. The use of a variety of alienation methods to break the inherent thinking presents a personalized visual effect, while expressing the design concept, so that the landscape installation becomes unique and individual, making the whole landscape space more attractive. The intervention of virtual image technology provides more means for the shaping of spatial differentiation. In landscape space, strong color contrast will produce certain visual signals to the public, and in the subjective world of human beings, color also has certain symbolic meaning, cultural meaning, warning meaning, etc., according to the designer’s concept of expression and the needs of the landscape space atmosphere for image color saturation, contrast, and the brightness of different collocation, to cause different color feelings and connotations to the public. In the 2013 Sydney Christmas light show, St. Mary’s Cathedral became the protagonist of the light show; the designer used wall projection technology in the church facade projection show, by changing the color of the church, so that the audience’s visual experience of the church and the inner feelings have changed; green projection makes the church warm and romantic, and when the dark red projection in the church appears, in the audience’s heart is a majestic green projection that makes the church warm and romantic, while when the dark red projection appears in the church, it creates a majestic visual and inner feeling in the audience. The meaning of color is complex and has different meanings in different regions, as shown in Figure 5.

The common coordinate systems used in visual-inertial navigation are the world coordinate system, the camera coordinate system, and the IMU coordinate system. The world coordinate system is a fixed reference coordinate system, the camera coordinate system is a coordinate system bound to the camera with the shooting viewpoint as the coordinate origin, and the IMU coordinate system is bound to the IMU device and is a moving coordinate system. Since the reconstruction method used in this paper is single-view reconstruction, the camera coordinate system is used as the reference coordinate system, and the world coordinate system is not considered, and only the camera coordinate system and the IMU coordinate system are fused to solve the motion trajectory and pose of the camera in the indoor scene. In pure visual SLAM, the camera coordinate system of the first frame is generally used as the world coordinate system. In this paper, the camera coordinate system of the first panoramic image is used as the reference coordinate system, and the 3D model reconstructed from subsequent panoramic images is converted to the reference coordinate system to realize the stitching of the model. The layout models reconstructed by single panoramic images are all in their respective camera coordinate systems as the reference coordinate systems, i.e., the coordinate systems of the models reconstructed by different panoramic images are relatively independent. To realize the stitching of multiple 3D models, it is necessary to obtain the position relationship between each model, i.e., the relative position relationship of each panoramic image shooting viewpoint. When the IMU device is fixed to the camera device, the relative position relationship between the two viewpoints, i.e., the relationship between the two camera coordinate systems, can be calculated from the IMU data when switching the shooting viewpoints. The relationship between the two coordinate systems can be represented by a rotation matrix and a translation matrix, and the camera model between the two coordinate systems can be converted to the same coordinate system after obtaining the camera model between the two coordinate systems to achieve model stitching. This is shown in Figure 6.

4.2. Image Virtual Reconstruction Implementation

The projection ratio of the ceiling area to the wall area on the panoramic graphic, C_h is the camera height, R^h is the room height, and an  _f, a_g denotes the vertical offset angle when the corresponding points of the ceiling area and the floor area are projected onto the sphere, respectively. The camera height and the height of the room scene affect the relationship between the projection of the ceiling area and the wall area on the image. The camera calibration ensures that the horizontal vanishing line in the scene is the x-axis of the image coordinate system and that the x-coordinate of the image coordinates of the two points corresponding to the floor and wall points in a real scene is the same when projected onto the panoramic image, and the relationship between the y-coordinates is determined by the projection scale. It mainly includes three interactive methods: translation, rotation, and zoom. In the actual interaction process, user interaction gestures are essentially two-dimensional (that is, the movement of the user’s finger on the display screen of the mobile device), but the MAR experience involves three dimensions in the real world. Therefore, this system chooses to limit the rotation of the virtual medical model to a single axis, and the translation range is limited to the plane where the virtual object was originally placed, to simplify the interaction between the user and the virtual object. This property is used to calculate the ground contour lines corresponding to the contour line segment of the ceiling area in the image. After determining the contour line segments of the ceiling and the floor, it is possible to find the ceiling area and the floor area in the image by the contour line segments. The closed-loop area enclosed by the contour line segments is the ceiling area, and in a panoramic image, the ceiling area contour line segments often span the entire image. When projected onto the sphere, it appears as a spherical region made up of several spherical triangles, each of which is enclosed by a line segment connecting the contour line segment to the upper vertex P of the sphere. When converted to the image coordinate system, the spherical triangles are represented as image regions above the contour line segments, combining this property. The ability to find the ceiling area in an image is shown in Figure 7.

This chapter focuses on a comprehensive verification of the entire experimental system, using a real-life scenario built on an experimental platform. The corresponding error analysis is given for some of the experiments covered in this paper. In the section on camera calibration, the internal parameters of the camera are calculated using two calibration plates. In the image correction section, the symmetry axis errors obtained for the six images are within 0.3 pixels. In addition, the two measurements of the long and short axes of the images before and after correction were compared with a difference of 0.0934 pixels and 0.1229 pixels, respectively, and the ratio of the two was compared with the ratio of the length and width of the real artifacts, and the maximum error value obtained was 0.0018, and the average error value was 0.0012, which has high accuracy and can be used for 3D plotting using the bus data obtained by the algorithm of this paper. Finally, the plotted model is imported into the real scene to realize the virtual reality alignment. The object of the experimental study is a rotating ceramic vase with a maximum circumference of 43 cm and a height of 23 cm, and the diameter of the belly diameter is 13.6873 cm. The ratio of the maximum diameter length to the height is calculated to be approximately equal to 0.5954, and the maximum error value is 0.0018, and the average error value is 0.0012 when compared with the data in the above table.