BGR + LBP two-parameter description method for registration feature points in aerial image of small unmanned aerial vehicle

Abstract

The invention discloses a BGR + LBP two-parameter description method for registration feature points in aerial images of a small unmanned aerial vehicle. The method comprises the following steps: acquiring BGR values of the feature points; constructing a feature point adjacent pixel 3 * 3 window; calculating a BGR mean value of the feature points and eight pixel points around the feature points; obtaining an adjacent pixel reference value of the feature point # imgabs0 #; and obtaining a BGR + LBP two-parameter description value of the feature point # imgabs1 #. According to the method, the color features of the feature points can be effectively displayed, strong constraint on the feature points can be achieved through the LBP values of the surrounding pixels, noise interference possibly existing around can be effectively reduced, and technical support can be provided for achieving efficient registration of the feature points in the later period and greatly reducing the calculation amount of registration of the feature points.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a BGR+LBP dual-parameter description method for registering feature points in aerial images taken by a small unmanned aerial vehicle. Background Art

[0002] In the 3D real scene modeling of construction engineering construction sites, the use of small drones to collect construction site information can efficiently and cost-effectively obtain a large number of construction site images and data. In the later 3D real scene modeling of construction sites, it is necessary to align these collected images and splice them into a whole. To achieve this goal, the first thing to do is to find the registration feature points in the adjacent images that need to be registered and describe the feature points. Among them, the description of feature points is a very important task. Effective feature point description is an important basis for achieving efficient and accurate image registration.

[0003] In terms of feature point description, the current method is to use a single parameter description of feature points using image grayscale values. This method is easily replaced by neighboring points with similar parameter values ​​in the feature point registration of adjacent images, resulting in a significant decrease in registration efficiency and accuracy. Summary of the invention

[0004] In view of the problems and defects existing in the above-mentioned prior art, the present invention proposes a BGR+LBP dual-parameter description method for registering feature points in aerial images of small unmanned aerial vehicles. This BGR+LBP dual-parameter feature point description method uses the BGR (Blue Green Red) value of the feature point itself as the main parameter and the local value LBP (Local Binary Pattern) formed by the 8 surrounding neighboring pixels as the auxiliary parameter.

[0005] A BGR+LBP dual-parameter description method for registering feature points in an aerial image taken by a small unmanned aerial vehicle, the method comprising the following steps: Get the BGR value of the feature point; Construct a 3×3 window of pixels adjacent to the feature point; Calculate the BGR mean of the feature point and the eight surrounding pixels; Get feature points The neighboring pixel reference value of Get feature points The BGR+LBP double parameter description value.

[0006] Further, obtaining the BGR value of the feature point specifically includes: using image analysis software to obtain any feature point in the image The BGR value of is the row and column coordinates of the point in the image.

[0007] Further, constructing a feature point neighboring pixel window specifically includes: clearly defining the selected feature point Then, take the feature point As the center, select a 3×3 feature point neighboring pixel window.

[0008] Further, calculating the BGR mean of the feature point and its surrounding eight pixels specifically includes: After obtaining the BGR of all pixels in the 3×3 neighboring pixel window of the feature point, calculate the BGR mean of each pixel , the algorithm is: ; ; ; ; ; ; ; ; ; Further, obtaining feature points The neighboring pixel reference value includes: The BGR mean Threshold , the BGR mean of the eight adjacent pixels and Make comparisons; If the BGR mean of the neighboring points is greater than the threshold, the pixel is assigned a value of 1; if the BGR mean of the neighboring points is less than the threshold, the pixel is assigned a value of 0.

[0009] Further, obtaining feature points The BGR+LBP double parameter description value includes: starting from the upper left corner, 8 reference values ​​are converted clockwise to generate 8-bit binary numbers, and then the binary numbers are converted to decimal values. This value is the value of the feature point is the LBP value of the center point In this way, the parameter value of BGR+LBP can form the feature point BGR+LBP dual parameter description value .

[0010] The present invention has the following beneficial technical effects: (1) The present invention can effectively display the color characteristics of feature points.

[0011] (2) The present invention can achieve strong constraints on feature points through the LBP values ​​of surrounding pixels, effectively reduce possible noise interference in the surroundings, and provide technical support for the later efficient registration of feature points and greatly reduce the calculation amount of feature point registration. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The present invention has the following accompanying drawings: Figure 1 This is a schematic diagram of a 3×3 neighboring pixel point window of a feature point of the present invention; Figure 2 Schematic diagram of the LBP method for obtaining feature points of the present invention; Figure 3 The present invention is a flowchart of the method of the present invention. DETAILED DESCRIPTION

[0013] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0014] The following is combined with Figure 1-3 The present invention is described in further detail.

[0015] An image is a picture composed of many pixels. Any pixel can be represented by an image grayscale value or a BGR value. When represented by grayscale values, the grayscale values ​​of the pixels in the image are usually marked with integers between 0 and 255, where 0 represents black, 255 represents white, and the values ​​in between represent different degrees of gray. This grayscale-based pixel description method is simple to calculate and has a small amount of data. However, since there are a large number of grayscale values ​​with equal values ​​in the image, they will be replaced by neighboring points with similar parameter values ​​in the feature point registration of adjacent images, reducing the efficiency and accuracy of image registration. More importantly, since the images of construction sites collected by drones are all color images, the single parameter description method using grayscale values ​​cannot effectively achieve the unique identification of each feature point.

[0016] To address this problem, in order to give each feature point a unique significance identifier, the present invention proposes a BGR+LBP dual-parameter feature point description method, which describes the feature point itself through the dual parameters consisting of the BGR value of the feature point and the local pattern binary grayscale value LBP formed by the 8 neighboring pixels around it, thereby achieving effective identification of each feature point and providing technical support for the unique mutual mapping of feature points in the subsequent image registration.

[0017] The specific embodiments are as follows: After image registration is carried out and the feature points of the image are obtained, it is necessary to describe each feature point in the image so that in the future adjacent image registration, mutual mapping can be carried out based on the uniqueness of the feature points to complete the registration and splicing of adjacent images. Using BGR+LBP dual parameters to describe and identify feature points includes the following five steps: Step 1: Get the BGR value of the feature point After obtaining the feature points, you can use image analysis software to obtain any feature point in the image The BGR value of is the row and column coordinates of the point in the image. For example, if the image analysis software OpenCV is used, the image can be read through the cv2.imread() function, and the BGR value of a specific pixel can be directly accessed through the coordinate value index. For example, image[x, y] can obtain the BGR value of the pixel at position (x, y), and image[x, y, 0], image[x, y, 1], and image[x, y, 2] can obtain the blue, green, and red values ​​of the pixel respectively. Since the image point is a color image, the BGR value of the image point contains three values: blue, green, and red. , such as (173,48,104).

[0018] Step 2: Construct a 3×3 window of pixels adjacent to the feature point Clarify the selected feature points Then, take the feature point As the center, select a 3×3 feature point neighboring pixel window, so that the window has 9 pixels, such as Figure 1 As shown. As the center, there are eight pixels around it, namely: ; Get feature points The same method as the BGR value can be used to obtain the BGR values ​​of the other eight points. The corresponding BGR values ​​are , , .

[0019] Step 3: Calculate the BGR mean of the feature point and its surrounding eight pixels After obtaining the BGR of all pixels in the 3×3 neighboring pixel window of the feature point, calculate the BGR mean of each pixel The algorithm is as follows: ; ; ; ; ; ; ; ; ; Step 4: Get feature points The neighboring pixel reference value After completing the BGR mean calculation of the feature point and its surrounding eight pixels, the feature point The BGR mean Threshold , the BGR mean of the surrounding 8 pixels and If the BGR mean of the neighboring points is greater than the threshold, the pixel is assigned a value of 1. If the BGR mean of the neighboring points is less than the threshold, the pixel is assigned a value of 0. In this way, the reference values ​​of the surrounding 8 adjacent pixels can be obtained. For example, the feature point The BGR mean is 90, and the average values ​​of the other eight pixels are 79, 112, 177, 170, 156, 89, 111, and 120 respectively, so the feature points can be obtained. The neighboring pixel reference values ​​are 0, 1, 1, 1, 1, 0, 1, 1. Figure 2 shown.

[0020] Step 5: Get feature points BGR+LBP dual parameter description value Getting feature points After comparing the neighboring pixel reference values, start from the upper left corner and generate 8 binary numbers from the 8 reference values ​​in a clockwise direction. Then convert the binary numbers into decimal values. This value is the value of the feature point. is the LBP value of the center point In this way, the parameter value of BGR+LBP can form the feature point BGR+LBP dual parameter description value For example, if the feature point The BGR value is (173, 48, 104), the generated binary value is 01111011, converted to decimal value , then the feature point The BGR+LBP dual parameter description value is (173, 48, 104, 123).

[0021] In order to clearly demonstrate the content and use method of the present invention, the present invention provides the following Figure 3 In the attached drawings shown, the blue framed portion is the content described in the present invention. Figure 3It is used to show the process and specific steps of implementing the BGR+LBP dual parameter description of image registration feature points in a more concise and clear manner.

[0022] The above embodiments describe the technical solutions of the present invention in detail. Obviously, the present invention is not limited to the described embodiments. Based on the embodiments of the present invention, people familiar with the technical field can also make various changes accordingly, but any changes that are equivalent or similar to the present invention belong to the scope of protection of the present invention.

[0023] 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 BGR+LBP dual-parameter description method for registering feature points in aerial images taken by small unmanned aerial vehicles, characterized in that: The method comprises the following steps: Get the BGR value of the feature point; Construct a 3×3 window of pixels adjacent to the feature point; Calculate the BGR mean of the feature point and the eight surrounding pixels; Get feature points The neighboring pixel reference value of Get feature points The BGR+LBP double parameter description value.

2. The method according to claim 1, characterized in that The method of obtaining the BGR value of the feature point specifically includes: using image analysis software to obtain any feature point in the image The BGR value of is the row and column coordinates of the feature point in the image.

3. The method according to claim 2, characterized in that The step of constructing a feature point neighboring pixel window specifically includes: determining the selected feature point Then, take the feature point As the center, select a 3×3 feature point neighboring pixel window.

4. The method according to claim 3, characterized in that The calculation of the BGR mean of the feature point and the eight surrounding pixels specifically includes: after obtaining the BGR of all pixels in the 3×3 neighboring pixel window of the feature point, calculating the BGR mean of each pixel , the algorithm is: ； ； ； ； ； ； ； ； 。 5. The method according to claim 4, characterized in that Acquiring feature points The neighboring pixel reference value includes: The BGR mean Threshold , the BGR mean of the eight adjacent pixels and Make comparisons; If the BGR mean of the neighboring points is greater than the threshold, the pixel is assigned a value of 1; if the BGR mean of the neighboring points is less than the threshold, the pixel is assigned a value of 0.

6. The method according to claim 5, characterized in that Acquiring feature points The BGR+LBP dual parameter description value specifically includes: starting from the upper left corner, 8 reference values ​​are converted clockwise to generate 8-bit binary numbers, and then the binary numbers are converted to decimal values. The decimal values ​​are based on the feature points. is the LBP value of the center point ; In this way, the parameter value of BGR+LBP constitutes the feature point BGR+LBP dual parameter description value .

Images