CN101499176B - Video game interface method - Google Patents

Abstract

The present invention is a video game interface method that realizes natural interaction with game scenes. Use the camera to collect the two-dimensional plane image of the aircraft with three marks. When the aircraft is in various movements, it can quickly and accurately identify the positions of the three marks of the aircraft, and through the method of coordinate transformation, the space can be calculated more accurately. Location. The steps include: obtaining data from a camera; obtaining three point sets through threshold segmentation; obtaining three point plane positions by a clustering method; obtaining spatial positions through coordinate conversion. The present invention is easy to realize and operate, so that game players have a newer interaction mode, which is different from the traditional mouse and keyboard interaction mode, so that players can obtain more immersion, and obtain more in the combination of physical objects and virtual objects in the game. much fun.

Description

A video game interface method

technical field

The invention belongs to the field of pattern recognition technology and digital interactive entertainment, and relates to a process of using video interactive technology to complete human-computer interactive entertainment.

Background technique

Human-computer interaction is an important part of game software, and its role is becoming more and more obvious with the development of computer hardware and software and the improvement of user needs. Gamers have long used traditional interaction methods such as mice, keyboards, joysticks and other dedicated gaming devices.

The purpose of the game itself is to make players live more entertaining and fresh. The traditional method is characterized by hand contact, and the process of human-computer interaction is completed in a "touchable" way. Players are already very familiar with it. Familiarity has become increasingly difficult to adapt to the complex and diverse needs of players. Therefore, more natural and intelligent interaction methods, such as sound, vision, motion and touch, etc., have been noticed by many people, and games based on video and audio have gradually appeared on the market, which have been widely welcomed.

In the field of digital interactive entertainment, video interaction technology has become a research hotspot in human-computer interaction in games in recent years, and cameras have become common equipment for personal computers. Currently, video-based interaction methods include: motion detection, face detection and expression recognition, gestures Detection, etc., but these methods are susceptible to interference such as changes in light, etc., and are still unstable.

Contents of the invention

The purpose of the present invention is to obtain the system of three-dimensional coordinates of three marked points on the player's aircraft held by the game player. A camera is used to collect a two-dimensional plane image of a toy plane with three light-emitting diodes, and the three-dimensional spatial positions of the three light-emitting diodes of the toy plane are recognized. The present invention is easy to realize and operate, so that game players have a newer interaction mode, which is different from the traditional mouse and keyboard interaction mode, so that players can obtain more immersion, and obtain more in the combination of physical objects and virtual objects in the game. much fun.

In order to achieve the above object, the steps of the video game interface method and system provided by the present invention include:

Step 1: Collect samples containing toy airplane images from video images;

Step 2: Perform threshold segmentation on pixel information in the video image to obtain the point set to be clustered in the next step;

Step 3: cluster the point set with a density-based clustering method (DBSCAN), obtain the set of three points formed by the three points in the video screen, and calculate the two-dimensional coordinates of the point in the form of the center of gravity;

Step 4: Using the method of three-point perspective, combined with the internal parameters of the camera and the external parameters of the aircraft model, the three-dimensional coordinates of the three points on the toy aircraft are obtained;

Step 5: Build a model with OpenGL for demonstration.

Further, the step of collecting a sample containing a toy airplane image from the video image includes:

Step 11: Obtain a 320×240 pixel image of each frame from the fisheye camera;

Step 12: Each obtained pixel is represented by 3 elements, that is, the R, G, B three-color model, and the range of each model is 0-255, respectively represented by a variable of type unsigned char in C++;

Step 13: Each frame of image obtained can store a total of 320×240×3 variables, and each pixel is represented by three components R, G, and B;

Step 14: Perform preprocessing on the obtained image of each frame, and remove the region where the grayscale deviates from the target set value.

Further, the step of performing threshold segmentation on pixel information in the video image to obtain the point set to be clustered in the next step includes:

Step 21: Carry out the preprocessing of the preprocessed image in the second step, and take out those isolated noise points and points that are continuously distributed and whose color information difference from the marked point is less than the set value;

Step 22: Select different threshold segmentation models according to the light intensity;

Step 23: Obtain different numbers of target points from different threshold segmentation models.

Further, the described method of clustering point sets based on density (DBSCAN) is used to obtain the set of three points formed by three points in the video screen, and the step of calculating the two-dimensional coordinates of the points in the form of center of gravity include:

Step 31: Change the parameter value of the model according to the number of points obtained in the previous step, and collect points again to ensure that the number of points is within the set range;

Step 32: converting the points obtained in step 31 into coordinate values in the plane video image;

Step 33: Initialize the density-based clustering method (DBSCAN) clustering;

Step 34: read in the coordinates of the two-dimensional points in step 33 sequentially, as the input and initial point of the clustering based on density clustering mode (DBSCAN);

Step 35: Preprocessing the read points;

Step 36: establish a relationship between points and points;

Step 37: Establish the core points of clustering and adjust according to the parameters;

Step 38: determine clustering;

Step 39: Gather the third group of points according to the clustering to obtain the center of gravity coordinates.

Further, the method of using the three-point perspective view, combined with the internal parameters of the camera and the external parameters of the aircraft model, the step of obtaining the three-dimensional coordinates of three points on the toy aircraft includes:

Step 41: Obtain the barycentric coordinates of the three clusters from step 39;

Step 42: Determine three coordinate systems, including image coordinate system, camera coordinate system and world coordinate system;

Step 43: use the tool ARToolKit to calibrate the camera, and obtain the internal parameters of the camera such as focal length, optical center and other information;

Step 44: Establish a conversion relationship between pixels and millimeter lengths;

Step 45: Obtain the camera space model by using the geometric positional relationship of the three coordinate systems;

Step 46: Obtain a three-point perspective view according to the spatial model of the camera;

Step 47: Obtain the external parameters of the toy plane, that is, the lengths of the three sides of the triangle formed between the three points, and convert it into pixel units;

Step 48: Calculate and obtain the coordinate conversion of the three-point perspective view;

Step 49: Obtain the three-dimensional coordinates of the three points from the three-point perspective view.

Further, the described steps of demonstrating with the OpenGL model include:

Step 51: Initialize the OpenGL model environment;

Step 52: correcting the obtained three-dimensional data, and normalizing it to data acceptable to OpenGL;

Step 53: building a triangular model in OpenGL through 3D data;

Step 54: The model in OpenGL moves in real time with the movement of the real toy airplane.

The present invention utilizes computer vision and image processing technology in a natural interactive game scene. Traditional interaction methods are characterized by hand contact, such as mouse and keyboard. Complete the process of human-computer interaction in a "touchable" way. With the development of computer technology, these traditional human-computer interaction technologies have become increasingly difficult to adapt to the complex and diverse needs of players. Gamers demand more natural and intelligent interaction methods, using computer vision to enable players to gain more immersion.

In addition, the present invention adopts the density-based DBSCAN algorithm, which can identify clusters of any shape, and only needs to manually set the initialization parameters, and the parameters can be manually set through experience, and have a relatively high value when there are many data efficiency. During the implementation of the algorithm, the search area of the core points is limited, which greatly reduces the search range and improves the search efficiency, making the clustering speed reach about 20ms, which meets the real-time requirements of video.

Description of drawings

Fig. 1 is the video effect figure that obtains at first of the present invention;

Fig. 2 is a video acquisition property diagram of the present invention;

Fig. 3 is the program structural diagram that the present invention is used in digital game;

Fig. 4 is the schematic diagram that the present invention is used in digital game;

Fig. 5 is the connection relationship of DBSCAN clustering process points;

Fig. 6 is a schematic diagram of core points and non-core points in the clustering process;

Figure 7 is a model of a three-point perspective (P3P).

Detailed ways

The present invention will be described in detail below in conjunction with the accompanying drawings, and the described embodiments are only intended to facilitate the understanding of the present invention, rather than limiting it in any way.

The operation process of a video game interface method and system will be further described below through an example.

All the codes in this example are written in C++ and run under Microsoft visual studio 2005 environment. The program structure diagram is shown in FIG. 3 , wherein 300 , 301 , 302 and 303 are components of the present invention, which can be connected to the game scene through the API interface.

As shown in FIG. 1 , the first step of the system shown is to acquire an image, such as 300 in FIG. 3 .

(1-1) Most of the information in digital image processing is two-dimensional information, and the amount of processed information is large. An image here is represented by a two-dimensional function f(x, y), where x, y are two-dimensional coordinates, and f(x, y) represents the color information of the point (x, y). The camera collects all the optical information in the lens from the space. After the information enters the computer, it is converted into a color model that meets computer standards, and then enters the program for digital image processing, and ensures the continuity and real-time performance of the video. A total of 320×240 pixels and 76800 pixels are processed for each pixel in the collected image. The initial effect of the collected video is shown in Figure 1. All subsequent operations and calculations of the project are based on the 320×240 pixels of each frame. The video capture properties shown in Figure 2.

(1-2) All colors in nature can be formed by combining the three primary colors of red, green and blue (R, G, B). According to the amount of red components, it can be artificially divided into 256 levels from 0 to 255. A grade of 0 means no red ingredients, and a grade of 255 means 100% red ingredients. Similarly, green and blue can also be divided into 256 levels. The video capture property shown in Figure 2 selects the RGB format. Such a method is easy to be processed by a computer, which improves convenience for the following stages, such as 301 in FIG. 3 .

(1-3) The next step is to identify the marks on the toy airplane. The markers currently used are red diodes about 1 cm in diameter, which can be better captured without emitting light. Its advantage is the spherical surface, which still has a good recognition effect when the angle changes. The area to be captured is the three circled places in Figure 1. It is necessary to perform 01 calibration on the image, that is, threshold segmentation. The data is taken to the next step to find clusters to determine the specific location of the luminescence in the 2D image. Assuming the original image f(x, y), find a suitable color image information in f(x, y) as the threshold t according to certain criteria, then the image g(x, y) segmented according to the above method can be obtained by the following formula means:

$\mathrm{<span\; class="notranslate">\; g</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Greater\; Equal;</span>\mathrm{<span\; class="notranslate">\; t</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; t</span>}\end{array}\right.$

Through repeated experiments, the r, g and b distributions represent the information of the three colors, and their ranges are all 0-255, where _r=r/(r+g+b), satisfying: _r>k And r+g+b>350. That is, the red information reaches a certain proportion of the total information. Here k is a variable parameter in the program, that is, the data will increase when the light is strong, so k should be reduced to reduce the number of collected points, and k will be increased correspondingly when the light is weak.

From the point set obtained in step (1-3), and clustering with this as input, you can find the positions of 3 markers in the planar video, such as the point set found in Figure 4, such as 302 in Figure 3 .

(2-1) DBSCAN (Density-Based Spatical Clustering of Application with Noise) is a kind of clustering algorithm. The purpose of clustering is to group data belonging to the same type and separate data of different types. DBSCAN is a density-based algorithm. , its basic idea is: for each object in a class, the number of objects contained in its given radius domain cannot be less than a given minimum number. Below are some definitions of the DBSCAN algorithm.

Definition: Eps-neighborhood of a point

The Eps-neighborhood of a point p is denoted by N _Eps (p), defined as:

N _Eps (p) = {q∈D|dist(p, q)≤Eps}

Definition: MinPts

When the number of points in the Eps-neighborhood of point p is greater than MinPts, the point is a core point. As shown in Figure 6, if the MinPts of A is 7, and Eps is the radius of the circle, then A is the core point, but B is not.

Definition: Clustering

Let D be a point in the database. A cluster C is a collection of points in D that are not empty and satisfy the following conditions:

p，q：如果p∈C而且q对于p来说是密度可达的，则q∈C。(最大化)(1)

p, q: If p ∈ C and q is density-reachable for p, then q ∈ C. (maximize)

p，q∈C：p对于q来说是密度连接的。(2)

p, q ∈ C: p is density-connected with respect to q.

As shown in Figure 4, the found points are immediately 3 clusters.

Definition: noise

i：p

C_i}。Let C ₁ , . . . , C _k be the clusters in the database D, i=1, . . . , k. We define noise as points in the database that do not belong to any cluster _Ci , i.e., noise = {p∈D|

i:p

C _i }.

(2-2) Step 1: Point reading and preprocessing. As mentioned in the point definition, each point has a member of the sSortedList class, which is a linked list of a series of points, arranged in order according to the size of the distance from the point. For example, there are the following points A (0, 0), B (1, 1), C (1, 1.5), D (2, 2), E (2.5, 0), then for point A, its sSortedList Members are shown in Figure 5.

(2-3) Step 2: Establish core points. The premise of completing this step is the association relationship of the points completed in step 1. First start from the first point, and then judge the members of the sSortedList class of each point. Select the MinPts node, if the distance of this node is less than Eps, set the core_tag value of this point to true. until all points are judged. Examples of core points and non-core points are shown in Figure 6.

Such a process can also be understood as drawing a circle with the point to be judged as the radius. If the number of points in the circle is not less than MinPts, it means that this point is the core point. The purpose of this step is to complete such preparations for each point. In the program, each point enters the class in the form of a linked list, and each node has members similar to the above figure. Sorting is done during insertion, and the time complexity is O(N ² ). If DBSCAN has to process N points in total, the length of the members of the sSortedList class of each point is N, that is to say, the entire class will have N×N nodes. If N is large, it will result in a large storage capacity and a long calculation time.

(2-4) After all core points are determined through the previous step, the density-based algorithm has a basis. The basic idea of DBSCAN is that for each object in a class, the range of its given radius cannot contain less than a given minimum number of objects. The clustering process steps are as follows:

First set a global cluster number current_class_id, initialized to 1. Then starting from the first point, if this point is a core point and has not been visited, set the class_id of the current point to the global cluster number current_class_id, set the classified_tag to true, and enter the core point clustering function CorePointCluster at the same time. After this process is completed, the global clustering number current_class_id is incremented until all point visits are completed. In this way, all clusters are found, that is, those with the same current_class_id number are the same cluster, and those with current_class_id 0 are noise points.

(2-5) Improvement of clustering algorithm In fact, the whole program uses only the nearest MinPts points around the point and the circle with Eps as the radius. Therefore, we only need to associate these points, and these points must be less than N. The time and space savings are considerable if the selection of MinPts and Eps is small. This meets the real-time requirements of the program.

After the clustering is completed, the plane coordinates of the three points are obtained, and the spatial coordinates of the three points are obtained through the three-point perspective (P3P) method, as shown in part 303 in FIG. 3 .

(3-1) Firstly, the camera must be calibrated. The purpose of calibration is to obtain the internal parameters of the camera, such as focal length, coordinates of light, etc.

(3-2) The P3P algorithm is nonlinear and has no exact solution, so an approximate algorithm is needed to obtain the solution. We use the algorithm for the two-point problem to solve the three-point problem. As shown in Figure 7, points m ₀ , m ₁ , m ₂ are the projections of points M ₀ , M ₁ , M ₂ in space on the image; line segment m ₀ m ₁ is the line segment M ₀ M ₁ in space in the image The projection on the line segment M ₀ M ₁ has _a length of D ₁ ; the line segment m ₀ m ₂ is the projection of the line segment M ₀ M ₂ in space on the image, and the length of the line segment M ₀ M ₂ is D ₂ ; Among the angles of the vertices, the angle M ₀ M ₁ M ₂ is known and is α. The shape of a triangle in space is completely determined by the three points m ₀ , m ₁ , m ₂ , the distance R ₀ from the extension line of the line segment Om ₀ to the point M ₀ and the angles θ ₁ and θ ₂ . By calculating R ₀ and θ ₁ , θ ₂ , the positions and angles of the three points M ₀ M ₁ M ₂ in the space can be determined.

Through the calculation of the two-point problem, according to the law of sine, we apply the line segment M ₀ M ₁ and the line segment M ₀ M ₂ , and the following relationship can be obtained:

$\frac{\mathrm{<span\; class="notranslate">\; sin</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; \&gamma;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="R"\; filenames="S2008100571820E00031.png"\; state="\{\{state\}\}">R</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&gamma;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}{{\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; ,</span>$ $\frac{\mathrm{<span\; class="notranslate">\; sin</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; \&gamma;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="R"\; filenames="S2008100571820E00031.png"\; state="\{\{state\}\}">R</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; sin</span>}{\mathrm{<span\; class="notranslate">\; \&gamma;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}{{\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; ,</span>$

In this way an equation for the angular relationship is found. Dividing the above two equations, shaving off R ₀ , we get:

$\frac{\mathrm{<span\; class="notranslate">\; sin</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; \&gamma;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; sin</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; \&gamma;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; K</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

Where K is obtained from the known variables γ ₁ , γ ₂ , D ₁ and D ₂ :

$\mathrm{<span\; class="notranslate">\; K</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; sin</span>}{\mathrm{<span\; class="notranslate">\; \&gamma;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; /</span>{\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}{\mathrm{<span\; class="notranslate">\; sin</span>}{\mathrm{<span\; class="notranslate">\; \&gamma;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; /</span>{\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}$

Another equation to obtain the relationship between angle θ ₁ and θ ₂ needs to be obtained from the angle α between the space line segment M ₀ M ₁ and M ₀ M ₂ :

cosα＝sinθ ₁ sinθ ₂ cosφ+cosθ ₁ cosθ ₂ (3)

Where φ is the angle between m ₀ m ₁ and m ₀ m ₂ in the image.

This way we get two equations (2) and (3) for _θ1 and _θ2 . Approximate solution to the equation:

Using the estimation method, the following two equations need to be solved:

$\frac{\mathrm{<span\; class="notranslate">\; sin</span>}{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}{\mathrm{<span\; class="notranslate">\; sin</span>}{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; K</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

sinθ ₁ sinθ ₂ cosφ+cosθ ₁ cosθ ₂ = cosα (5)

Put equation (4) to get $\sin {\mathrm{\&theta;}}_2=\frac{\sin {\mathrm{\&theta;}}_1}{K},$ Therefore, sinθ ₂ can be eliminated. Move cosθ ₁ cosθ ₂ in equation (5) to the right, and square both sides to get,

cos ² α+sin ² θ ₁ sin ² θ ₂ cos ² φ-2cos α sin θ ₁ sin θ ₂ cos φ＝cos ² θ ₁ cos ² θ ₂ will $\sin {\mathrm{\&theta;}}_2=\frac{\sin {\mathrm{\&theta;}}_1}{K}$ Putting it into the above formula will get an equation about sin ² θ ₁ , let X ₁ = sin ² θ ₁ , and simplify it to get,

${\mathrm{<span\; class="notranslate">\; sin</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; \&phi;</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; K</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; sin</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}$

The equation can have two solutions for X ₁ , but only the smaller one is less than 1, possibly a sinusoidal value. Therefore, the solution for sinθ ₁ can be obtained from the subquadratic equation:

$\mathrm{<span\; class="notranslate">\; sin</span>}{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; [</span>\frac{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; -</span>\sqrt{\mathrm{<span\; class="notranslate">\; \&Delta;</span>}}}{\mathrm{<span\; class="notranslate">\; 2</span>}{\mathrm{<span\; class="notranslate">\; sin</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; \&phi;</span>}}<span\; class="notranslate">\; ]</span>}^{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}}$

in,

$\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; \&Delta;</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; B</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 4</span>}{\mathrm{<span\; class="notranslate">\; K</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; sin</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}{\mathrm{<span\; class="notranslate">\; sin</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; \&phi;</span>}$

The value of sin θ ₁ can be obtained, and two values of θ ₁ can be obtained, which are related to the direction of the line segment M ₀ M ₁ . According to the formula (X), it can be seen that:

So there is only one final solution R ₀ (the length of line segment OM ₀ ).

The distance R ₀ from the center of the camera is obtained, as shown in Figure 7, and the angle between each point and the optical center of the camera is known, and the depth information of the point can be calculated according to the triangular relationship, combined with the coordinates of the plane, we can get specific location in space.

The present invention utilizes computer vision and image processing technology in a natural interactive game scene. Traditional interaction methods are characterized by hand contact, such as mouse and keyboard. Complete the process of human-computer interaction in a "touchable" way. With the development of computer technology, these traditional human-computer interaction technologies have become increasingly difficult to adapt to the complex and diverse needs of players. Gamers demand more natural and intelligent interaction methods, using computer vision to enable players to gain more immersion.

In addition, the present invention adopts the density-based DBSCAN algorithm, which can identify clusters of any shape, and only needs to manually set the initialization parameters, and the parameters can be manually set through experience, and have a relatively high value when there are many data efficiency. During the implementation of the algorithm, the search area of the core points is limited, which greatly reduces the search range and improves the search efficiency, making the clustering speed reach about 20ms, which meets the real-time requirements of video.

The above is only a specific implementation mode in the present invention, but the scope of protection of the present invention is not limited thereto. Anyone familiar with the technology can understand the conceivable transformation or replacement within the technical scope disclosed in the present invention. All should be covered within the scope of the present invention, therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims (4)

1. A video game interface method, characterized in that: comprising steps: Step 1: Collect samples containing toy airplane images from video images; Step 2: Perform threshold segmentation on each pixel information in the video image to obtain the point set to be clustered in the next step; Step 3: Use the density-based clustering method DBSCAN to cluster the point set to obtain the set of three points formed by the three points in the video screen, and calculate the two-dimensional coordinates of the point in the form of the center of gravity; Step 4: Using the method of three-point perspective, combined with the internal parameters of the camera and the external parameters of the aircraft model, the three-dimensional coordinates of the three points on the toy aircraft are obtained; Step 5: Build a model with OpenGL for demonstration. 2. The video game interface method according to claim 1, wherein said step of collecting samples comprising images of toy airplanes from video images comprises: Step 11: Obtain a 320×240 pixel image of each frame from the fisheye camera; Step 12: Each obtained pixel is represented by 3 elements, that is, the R, G, B three-color model, and the range of each model is 0-255, respectively represented by a variable of type unsigned char in C++; Step 13: A total of 320×240×3 variables are stored for each frame of image obtained, and each pixel is represented by three components R, G, and B; Step 14: Perform preprocessing on the obtained image of each frame, and remove the region where the grayscale deviates from the target set value. 3. The video game interface method according to claim 2, wherein the step of performing threshold segmentation on each pixel information in the video image to obtain the point set to be clustered in the next step comprises: Step 21: Carry out the preprocessing of the preprocessed image in the second step, and take out those isolated noise points and points that are continuously distributed and whose color information difference from the marked point is less than the set value; Step 22: Select different threshold segmentation models according to the light intensity; Step 23: Obtain different numbers of target points from different threshold segmentation models. 4. video game interface method according to claim 1, is characterized in that, the described step of demonstrating with OpenGL model comprises: Step 51: Initialize the OpenGL model environment; Step 52: correcting the obtained three-dimensional data, and normalizing it to data acceptable to OpenGL; Step 53: building a triangular model in OpenGL through 3D data; Step 54: The model in OpenGL moves in real time with the movement of the real toy airplane.

Images