Aerial image feature point display enhancement method based on BGR value conversion technology

Abstract

The invention relates to an aerial image feature point display enhancement method based on a BGR value conversion technology. According to the method, enhanced display of the feature points in the adjacent image in the image is realized through a technology of converting the BGR values of the local pixel points of the feature points in the adjacent image, so that the feature points have remarkable features and can be highlighted in the image, and rapid identification of modeling software is facilitated. According to the method, BGR + LBP double parameters for describing the feature points are reserved, the feature points can be displayed in the image, the identification performance of the feature points is improved, no adverse effect is generated on the original image effect, and the method has very high practical value for improving the image registration accuracy and carrying out later image precision inspection.

Description

Technical Field

[0001] The invention relates to the field of engineering information technology, in particular to an aerial image feature point display enhancement method based on BGR value conversion technology. Background Art

[0002] In the 3D real-scene modeling of construction sites, small drones can be used to efficiently and cheaply obtain a large number of construction site images and data. However, in order to stitch these large numbers of images together, it is necessary to find the feature points of adjacent image registration in the aerial images in the 3D real-scene modeling of the construction site, and accurately describe and display these feature points to meet the needs of image registration and stitching.

[0003] At present, the current practice to complete this work is: first input the construction site pictures collected by drones into the modeling software, then complete the neighboring image feature point extraction according to the feature point extraction algorithm of different modeling software, and use the image grayscale value to describe the feature points, and then carry out neighboring image registration work on this basis.

[0004] However, due to the presence of a large number of similar feature points in neighboring images, the feature points extracted by the modeling software do not have prominent features, and the grayscale values ​​of the feature points are also very similar. The extracted feature points remain the same in the image. Therefore, in the feature point registration of neighboring images, they are easily replaced by neighboring points with similar grayscale values, making the image registration work slow, inefficient, and the registration effect poor. Summary of the invention

[0005] In view of the defects existing in the prior art, the purpose of the present invention is to provide a method for enhancing the display of feature points of aerial images based on BGR value conversion technology. The method is based on the BGR (Blue Green Red) + LBP (Local Binary Pattern) dual-parameter feature point description method, and through the BGR value conversion technology of the local pixel points of the feature points in the adjacent images, the enhanced display of the feature points in the above adjacent images is realized, so that they have significant features and can be highlighted in the image, which is convenient for rapid recognition of modeling software, and provides technical support for improving the efficiency and accuracy of image registration.

[0006] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0007] A method for enhancing the display of feature points of aerial images based on BGR value conversion technology, characterized in that it comprises the following steps:

[0008] Step 1: Select feature points p of neighboring images in the aerial image ij , perform BGR+LBP dual parameter description on the feature point;

[0009] Step 2: Take the feature point p selected in step 1 ij As the center, select the neighboring pixel point domain;

[0010] Step 3: Get the feature point p obtained in step 2 ij The BGR values ​​of all pixels in the neighboring pixel domain (c bij ,c gij ,c rij );

[0011] Step 4: construct a ij (c b ,c g ,c r ) is the BGR value expansion matrix CP centered ij , where (c b ,c g ,c r ) is the feature point p ij BGR value of

[0012] Step 5, select the pixel BGR conversion standard value;

[0013] Step 6, p ij The BGR values ​​of some pixels in the neighboring pixel domain are converted according to the pixel BGR conversion standard value obtained in step 5 to obtain feature points for display enhancement.

[0014] On the basis of the above scheme,

[0015] The selection of the neighboring pixel domain described in step 2 is specifically as follows: taking the feature point p ij As the center, select the 9*9 feature point neighboring pixel domain.

[0016] On the basis of the above scheme,

[0017] The BGR value matrix CP described in step 4 ij Specifically:

[0018]

[0019] On the basis of the above scheme,

[0020] The pixel BGR conversion standard value in step 5 is: ij The RGB monochrome value with high background distinction in the aerial image.

[0021] On the basis of the above scheme,

[0022] The RGB monochrome values ​​include: blue with a BGR value of (0,0,255), green with a BGR value of (0,255,0), and red with a BGR value of (255,0,0).

[0023] On the basis of the above scheme,

[0024] Step 6 described in p ij The BGR values ​​of some pixels in the neighboring pixel domain are converted according to the pixel BGR conversion standard value, specifically:

[0025] Ensure that the feature point p ij Except for the BGR values ​​of the other 8 neighboring pixels in the surrounding 3*3 pixel window, the BGR values ​​of the remaining pixels in the neighboring pixel domain are converted to the pixel BGR conversion standard value.

[0026] The method for enhancing the display of feature points of aerial images based on BGR value conversion technology of the present invention has the following beneficial effects:

[0027] The feature point display enhancement method realizes the enhanced display of the feature points in the above-mentioned adjacent images by converting the BGR value of the local pixel points of the feature points in the adjacent images, so that the feature points have significant features and can be highlighted in the image, which is convenient for the rapid recognition of the modeling software. This method can be used to write a feature point highlighting and enhancement analysis program, and then the program can be implanted into the current three-dimensional modeling software for engineering construction sites, so that the acquired feature points can be displayed and enhanced. Since this new feature point highlighting and enhancement technology not only retains the BGR+LBP dual parameters describing the feature points, but also makes the feature points appear in the image, it improves the identification of the feature points, and does not have an adverse effect on the original image effect. Therefore, it has a high practical value for improving the accuracy of image registration and conducting later image accuracy inspections. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The present invention has the following accompanying drawings:

[0029] Figure 1 This is a schematic diagram of the pixel area adjacent to the feature point described in Example 1;

[0030] Figure 2 This is a comparison diagram of the effects before and after the feature point display enhancement described in Example 1;

[0031] Figure 3 This is a technical solution flow chart of the feature point display enhancement method described in the present invention. DETAILED DESCRIPTION

[0032] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0033] An image is a picture composed of many pixels. Any pixel can be represented by an image grayscale value or a BGR value. When described by a single parameter, the grayscale value of the pixel in the image is represented by an integer between 0 and 255, where 0 represents black, 255 represents white, and the values ​​in between represent different degrees of gray. When described by BGR values, each pixel p ij The BGR value is composed of three parameter values ​​including Blue, Green and Red (c bij ,c gij ,c rij ), the minimum values ​​of Blue, Green, and Red are all 0, and the maximum values ​​are all 255, for example (173, 48, 104). Among them, ij is the row and column coordinates of the point in the image.

[0034] Compared with the method of using image grayscale values ​​to describe feature points, the BGR+LBP dual-parameter feature point description method is more effective and makes the feature points more unique identifiers. In view of the advantages of the BGR+LBP dual-parameter feature point description method, after carrying out the neighboring image registration work and obtaining the feature points of the neighboring images, the BGR+LBP dual-parameter feature point description method can be used to complete the description of the feature points. On this basis, the method of the present invention can be used to perform enhanced display operations on the feature points, so that the feature points have prominent features and stand out in the image while maintaining the unique identifier.

[0035] The feature point display enhancement method of the present invention is as follows Figure 3 As shown, it includes the following 5 steps.

[0036] (1) Step 1: Constructing the neighboring pixel domain of feature points

[0037] Clearly select the feature point p ij Then, take the feature point p ij As the center, select a 9X9 feature point neighboring pixel point domain. In this way, the feature point domain has 81 pixels. ij As the center, the pixel point at the upper left corner of the feature point domain is p i-4,j-4 , the pixel in the lower right corner is p i+4,j+4 ,like Figure 1 shown.

[0038] (2) Step 2: Obtain feature point p ij BGR value of the pixel in the neighboring pixel domain

[0039] At the clear feature point p ij After the neighboring pixel domain is obtained, a software is selected from a large number of image analysis software to obtain the BGR values ​​of all pixels in the domain (c bij ,c gij ,crij For example, using OpenCV software, you can use the cv2.imread() function to read an image in OpenCV software, and directly access the BGR value of a specific pixel through the coordinate value index. For example, image[i,j] can get the BGR value of the pixel at position [i,j], and image[i,j,0], image[i,j,1], and image[i,j,2] can get the blue, green, and red values ​​of the pixel respectively (c b ,c g ,c r ). (3) Step 3: Construct a ij The BGR value expansion matrix CP is centered ij

[0040] The feature point p ij The 9X9 feature point neighboring pixel domain has a total of 81 pixels, and the pixel in the upper left corner is p i-4,j-4 , the pixel in the lower right corner is p i+4,j+4 . Using the BGR values ​​of 81 pixels, we can construct a pixel ij (c b ,c g ,c r ) is the feature point expansion matrix CP ij .

[0041]

[0042] (4) Step 4: Select the pixel BGR conversion standard value

[0043] The pixel BGR value of the image (c b ,c g ,c r ) has three single colors, (0,0,255) for blue, (0,255,0) for green, and (255,0,0) for red. These three colors have obvious color highlighting features in the image. Therefore, if the BGR value of a pixel is changed to any of the three single colors (for example, change (173,48,104) to (0,0,255)), the pixel will stand out in the image. Since (255,0,0) red is more obvious in the image, (255,0,0) is generally selected as the conversion standard. If the image to be registered has a dark background, the BGR value can also be converted to (255,255,255) white or other color values ​​as the conversion standard.

[0044] (5) Step 5: p ij The BGR values ​​of the pixels in the neighboring pixel domain are converted

[0045] After selecting the BGR value conversion standard value, in addition to keeping the feature point pij The BGR values ​​of the other 8 neighboring pixels in the 3X3 pixel window around it remain unchanged (the purpose is to keep the dual parameter description features of BGR+LBP of the feature point unchanged). The feature point p is converted into ij The BGR values ​​of the remaining 72 pixels in the extended matrix are converted to the standard values ​​selected in step 4. For example, if the BGR values ​​of the remaining 72 pixels are converted using OpenCV software, the matrix CP can be expanded based on the feature points. ij , which is accomplished by assigning the language image[x,y,0]=new_red.

[0046] After completing the BGR value conversion of 72 pixels in this way, the feature point p can be placed in the image. ij Stand out. Figure 2 Taking the shown example, in order to highlight a feature point in the image, the BGR values ​​of the 72 pixels in the pixel domain adjacent to the feature point can all be converted to (255,0,0) values. After the conversion, you can see that the feature point is highlighted in the original image and is very easy to identify. This can provide technical support for modeling software to quickly and accurately identify feature points and improve image registration efficiency and accuracy.

[0047] The contents not described in detail in this specification belong to the prior art and professional common sense known to professional and technical personnel in this field.

Claims

1. A method for enhancing the display of feature points of aerial images based on BGR value conversion technology, characterized in that: The steps include: Step 1: Select the feature point p of the adjacent image in the aerial image ij , perform BGR+LBP dual parameter description on the feature point; Step 2: Take the feature point p selected in step 1 ij As the center, select the neighboring pixel point domain; Step 3: Get the feature point p obtained in step 2 ij The BGR values ​​of all pixels in the neighboring pixel domain (c bij ,c gij ,c rij ); Step 4: construct a ij (c b ,c g ,c r ) is the BGR value expansion matrix CP centered ij , where (c b ,c g ,c r ) is the feature point p ij BGR value of Step 5, select the pixel BGR conversion standard value; Step 6, p ij The BGR values ​​of some pixels in the neighboring pixel domain are converted according to the pixel BGR conversion standard value obtained in step 5 to obtain feature points for display enhancement.

2. The method for enhancing the display of feature points of aerial images based on BGR value conversion technology according to claim 1, characterized in that: The selection of the neighboring pixel domain described in step 2 is specifically as follows: taking the feature point p ij As the center, select a 9*9 feature point neighboring pixel domain.

3. The method for enhancing the display of feature points of aerial images based on BGR value conversion technology according to claim 2, characterized in that: The BGR value matrix CP described in step 4 ij Specifically:

4. The method for enhancing the display of feature points of aerial images based on BGR value conversion technology according to claim 1, characterized in that: The pixel BGR conversion standard value in step 5 is: ij The RGB monochrome value with high distinction from the background of the aerial photo.

5. The method for enhancing the display of feature points of aerial images based on BGR value conversion technology according to claim 4, characterized in that: The RGB monochrome values ​​include: blue with a BGR value of (0,0,255), green with a BGR value of (0,255,0), red with a BGR value of (255,0,0), and white with a BGR value of (255,255,255).

6. The method for enhancing the display of feature points of aerial images based on BGR value conversion technology according to claim 1, characterized in that: Step 6 described in p ij The BGR values ​​of some pixels in the neighboring pixel domain are converted according to the pixel BGR conversion standard value, specifically: Ensure that the feature point p ij Except for the BGR values ​​of the other 8 neighboring pixels in the surrounding 3*3 pixel window, the BGR values ​​of the remaining pixels in the neighboring pixel domain are converted to the pixel BGR conversion standard value.

Images