A ship AR display method for ship image recognition and positioning information

Abstract

The present invention relates to a ship AR display method based on ship image recognition and positioning information, the method comprising the following steps: step 1: obtaining ship positioning information and related video surveillance, and converting the positioning information into point coordinates of image information through coordinate system conversion; step 2: judging the ship in the image in the video surveillance through artificial intelligence, and obtaining the rectangular coordinates of the ship; step 3: associating and matching the ship point coordinates with the rectangular coordinates; step 4: labeling the ship image in the image with information such as the name according to the matching result, and performing AR display.

Description

Technical Field

[0001] The present invention relates to a computational analysis method, and in particular to a method for ship AR display of ship image recognition and positioning information, belonging to the technical field of image recognition. Background Art

[0002] With the rapid development of information technology, artificial intelligence, GPS positioning, AR display, and other technologies, related applications are gradually advancing. However, the positioning of inland vessels is generally transmitted every 30 seconds to 5 minutes, and the video surveillance range is generally around 300 to 600 meters. Directly converting positioning coordinates for AR display can easily result in inconsistencies between the ship image and the ship's position, as well as slow movement. To address this specific issue of ship video AR display, ship image recognition technology is used to correct the position, achieving linkage between positioning and image, and linked movement. Summary of the Invention

[0003] The present invention addresses the problems existing in the prior art and provides a ship AR display method for ship image recognition and positioning information. It uses artificial intelligence, GPS positioning, and AR display to process the overlay display of ship identity in ship images, effectively solving problems such as inconsistency between ship images and ship positioning marks, and slow movement of ship positioning marks.

[0004] To achieve the above objectives, the technical solution of the present invention is as follows: a method for ship image recognition and ship positioning information AR display, the method comprising the following steps:

[0005] Step 1: Obtain ship positioning information and related video surveillance. Because the map space is relatively small, we use the Mercator projection method to ensure data accuracy. The Mercator projection accuracy is "meter". Without considering distortion, the formula can be used X, X, Y, and Z are the Mercator coordinates of the ship's geographic location, thus obtaining the video image coordinates u and v. In addition, considering that the ship height Z in the same picture is basically a fixed value, the formula can be simplified to u = a1*X+b1*Y+C1, v = a2*X+b2*y+C2;

[0006] Step 2: Use artificial intelligence to identify the ship in the video surveillance image, use the artificial intelligence yolo algorithm model to perform image detection, and obtain the rectangular coordinates of the ship [x0, y0, x1, y1];

[0007] Step 3: Associate and match the coordinates of the ship point with the coordinates of the rectangle, and calculate the coordinates of the ship point [u, v] and the center point of the rectangle coordinates.

[0008] The distance [(x0+x1) / 2,(y0+y1) / 2], that is According to the ship's 30-50-meter voyage in 30 seconds and the length of the inland river ship is 30-60 meters, the length between the point coordinates and the center of the rectangle is less than When the potential matching relationship is confirmed, after completing the distance calculation and judgment of all point coordinates and rectangle coordinates, (1) prioritize one rectangle corresponding to one point and confirm the matching; (2) then process the situation when multiple points are potentially matched with a matrix coordinate, match the point with the smallest distance with it, and then check whether there are any matrix coordinates adjacent to the matrix coordinate that do not have potential matching point coordinates, calculate the distance between the unmatched matrix coordinates and the point coordinates nearby, and if the distance is less than 2.5 rectangle lengths, then match; (3) finally, do not match the remaining unmatched rectangle coordinates and point coordinates;

[0009] Step 4: Label the ship image in the image with information such as name based on the matching results, and display it in AR.

[0010] Compared with the prior art, the present invention has the following advantages:

[0011] 1) Solved the problem of the actual image of the ship and the ship positioning AR mark not overlapping due to the long time it takes to send positioning data;

[0012] 2) Solved the problem that the ship positioning AR marker could not move in real time due to the long transmission time of positioning data;

[0013] 3) Solved the problem of inconsistency between the actual image of a ship and the AR marker of the ship's location because some ships did not have positioning data and were close to ships with positioning data;

[0014] 4) By displaying the ship's identity in real time through video surveillance, it facilitates water supervision work, allowing supervisors to pay real-time attention to the specific conditions of hazardous chemical transport ships and illegal ships, and provide real-time communication and guidance;

[0015] 5) This algorithm integrates image detection, GPS positioning, AR display and other technologies in artificial intelligence, and can well adapt to the long time it takes to send positioning data of inland ships, the congestion and overlap of inland ships, etc. Compared with other algorithms,

[0016] This method has better targeting, high efficiency and wide application range. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a schematic diagram of the overall steps of Example 1 of the present invention;

[0018] Figure 2 This is a schematic diagram of the final effect of Example 1 of the present invention; DETAILED DESCRIPTION

[0019] In order to deepen the understanding of the present invention, this embodiment is described in detail below with reference to the accompanying drawings.

[0020] Example 1: See Figure 1 As shown, a ship AR display method for ship image recognition and positioning information includes the following steps:

[0021] Step 1: Obtain ship positioning information (Mercator coordinates) through ship positioning equipment, including: ship A [13209234.54, 3758824.3], ship B [13209197, 3758826.7]. Use the approximate calculation formula u = a1*X+b1*Y+C1, v = a2*X+b2*y+C2, where a1: 65.31879, b1: -49.0311659, c1: 678507843.8; a2: -13.106514, b2: -22.22368, c2: 256663062.7, to obtain the image coordinates of ship A [3835, 1137] and ship B [1266, 1576].

[0022] Step 2: Using image recognition technology, obtain the image rectangle coordinates of ship X [2425,853,5461,1441]; obtain the image rectangle coordinates of ship Y [2737,1333,0,1841];

[0023] Step 3: Perform a preliminary match between the positioning ship image position and the image recognition ship image position:

[0024] Step 3.1 Calculate the center point of the image recognition ship, the ship X center point is [3943, 1147], the ship Y center point is [1368.5, 1587];

[0025] Step 3.1 Calculate the length of the image recognition ship image, the length of the ship X rectangle is 4316, the length of the ship Y rectangle is 2315: Step 3.2 Calculate the distance between the positioning ship image position and the image recognition ship image position and match them:

[0026] (1) The center distance between ship A and ship X: 108 is less than 4316, A and X are matched

[0027] (2) The center distance between ship A and ship Y: 2711 is greater than 2315, A and Y do not match

[0028] (3) The center distance between ship B and ship X: 2508 is less than 4316, B and X are matched

[0029] (4) The center distance between ship B and ship Y: 104 is less than 2315, B and Y are matched

[0030] Step 4: One-to-one matching is performed between the image position of the located ship and the image position of the image-recognized ship. (1) The number of located ships matching with ship X is: ship A: 108 < ship B: 2508, so ship X matches ship A; (2) The number of located ships matching with ship Y is: ship B, so ship Y matches ship B.

[0031] Step 5: Label the ship image in the image with its name and other information based on the matching results, and display it in AR. Figure 2 .

Claims

1. A ship AR display method based on ship image recognition and positioning information, characterized in that: The method comprises the following steps: Step 1: Obtain ship positioning information and related video surveillance, and convert the positioning information into point coordinates of image information through coordinate system conversion; Step 2: Use artificial intelligence to identify the ship in the video surveillance image and obtain the rectangular coordinates of the ship; Step 3: Associate and match the coordinates of the ship point with the coordinates of the rectangle; Step 4: Label the ship image in the image with its name and display it in AR based on the matching results; Step 5: Before the next positioning information is sent, the ship in the video surveillance image is judged by artificial intelligence and the linear filtering algorithm to complete the continuous labeling of the same object and AR display; Among them, step 1: convert the positioning information into the point coordinates of the image information through coordinate system conversion to obtain the ship positioning information and related video surveillance. Because the map space range is relatively small, we use the Mercator projection method to ensure data accuracy. The Mercator projection accuracy is "meter". Without considering distortion and the basic difference in ship height in the same picture is small, the simplified formula u=a1*Xw+b1*Yw+C1,v=a2*Xw+b2*Yw+C2 is used to convert the ship's geographic coordinates (Xw, Yw) in the Mercator coordinate system into video image coordinates (u, v).

2. The ship AR display method for ship image recognition and positioning information according to claim 1 is characterized in that: Step 2: Use artificial intelligence to determine the ship in the video surveillance image and obtain the rectangular coordinates of the ship: Use the Yolo algorithm model to perform image detection and obtain the ship's position information [x0, y0, x1, y1].

3. The ship AR display method for ship image recognition and positioning information according to claim 1, characterized in that: Step 3: Associate and match the ship point coordinates with the rectangle coordinates: (1) Based on the fact that a ship travels 30-50 meters in 30 seconds and the length of an inland ship is 30-60 meters, the length of the point coordinates and the center of the rectangle should be within the range of 2 times the rectangle length. Priority is given to matching the case where there is only one point within this range; (2) When multiple points are within this range, they are marked according to the closest distance. At the same time, the remaining points are calculated to see if there are any unmatched ship rectangle coordinates nearby. If the distance between the coordinates is less than 2.5 rectangle lengths, they are matched; (3) No matching is performed on the remaining unmatched rectangle coordinates and point coordinates.

4. The ship AR display method for ship image recognition and positioning information according to claim 1, characterized in that: Step 4: Label the ship image in the image with its name based on the matching results, and perform AR display: Display the matched ship in a video overlay based on the ship identity in the positioning information and the ship's rectangular coordinates.

Images