Method for acquiring coordinates of feature points of target object when flying target object lands for first time
By obtaining the point coordinates of the parabolic line virtual line before and after the first landing of the flight target, the problem of large measurement errors in the existing technology is solved, and higher accuracy is achieved.
Patent Information
- Application Number
- CN202411362529.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-27
- Publication Date
- 2025-01-28
AI Technical Summary
Existing machine vision-based ranging technology is difficult to accurately obtain the coordinates of the target's characteristic point when the target object lands for the first time, resulting in large measurement errors.
Multi -frame images are collected by fixed -point to obtain a parabolic line virtual line before and after the target landing.
The measurement error is reduced, and the accuracy of the coordinate acquisition is improved when the target landed for the first time.
Smart Images
Figure CN119374479A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of distance measurement based on machine vision, and in particular to a method for acquiring coordinates of feature points of a flying target when the flying target lands for the first time. Background Art
[0002] When a flying target flies along a parabolic trajectory and finally lands, it is necessary to measure the flying distance of the flying target, such as the flight distance measurement of a solid ball. The key and difficulty of measuring the flying distance of a flying target lies in determining the first landing point of the target. In actual operation, ranging based on machine vision is becoming more and more mature. By collecting multiple frames of images, the key frame of the first landing of the target is captured, and then the actual coordinates of the first landing point are calculated through image analysis. If the key frame can be accurately captured and judged every time, the result is relatively accurate, but it is well known that due to the time interval between consecutive frames, the key frame will not be captured every time. To solve this problem, the existing technology is that if the key frame is not collected, the frame closest to the key frame is selected as a substitute. This method has a large error, and the error is also greatly related to the flight speed and acceleration of the flying target. It is difficult to control and needs to be improved urgently. Summary of the invention
[0003] In order to solve the technical problem of insufficient accuracy in the background technology, the present invention proposes a method for obtaining the coordinates of the characteristic points of a flying target when the flying target lands for the first time, comprising:
[0004] Collect multiple frames of images at a fixed point, wherein the multiple frames of images include multiple frames of first-type images before the target with target characteristics lands for the first time and one or more frames of second-type images after the target lands for the first time;
[0005] Acquire a first virtual line according to multiple frames of the first type of image, the first virtual line passing through the target feature point when the target first lands within an allowable error range, and acquire a second virtual line according to one or more frames of the second type of image, the second virtual line passing through the target feature point when the target first lands within an allowable error range;
[0006] The coordinates of the intersection of the first virtual line and the second virtual line are obtained as the coordinates of the characteristic point of the target object when the target object first lands.
[0007] The first virtual line is a parabola followed by the feature points of the target object before the target object first lands on the ground, which is fitted based on multiple frames of the first type of images.
[0008] The step of obtaining the first virtual line specifically includes:
[0009] Acquire a first image and a second image from the plurality of frames of first type images;
[0010] Acquire the first coordinates of the key points of the target object at the time of capturing the first image, and analyze the second coordinates of the key points of the target object at the time of capturing the second image according to the second image;
[0011] The straight line passing through the first coordinate and the second coordinate is obtained as the first virtual line.
[0012] It is characterized in that the first image is specifically the last frame of image collected before the target object first lands on the ground, or the second image and the first image are continuous frames.
[0013] Acquiring the first image specifically includes:
[0014] Sequentially obtain the height of the key point of the target object corresponding to the acquisition moment of each frame of image;
[0015] Calculate the height difference between the key points of two targets corresponding to two adjacent frames of images in sequence;
[0016] When the height difference is less than a preset value, the previous frame of two adjacent frames of images corresponding to the height difference is defined as the first image.
[0017] The second virtual line is a parabola followed by the feature points of the target object after the target object first lands, which is fitted based on multiple frames of the second type of images.
[0018] The second virtual line is a horizontal straight line, and the second virtual line passes through a target feature point of a target located on the ground.
[0019] The step of obtaining the second virtual line specifically includes:
[0020] Acquire a third image and a fourth image from the plurality of frames of the second type of images;
[0021] Acquire the third coordinates of the key points of the target object at the time of capturing the third image, and analyze the fourth coordinates of the key points of the target object at the time of capturing the fourth image according to the fourth image;
[0022] The straight line passing through the third coordinate and the fourth coordinate is obtained as the second virtual line.
[0023] The third image is specifically the first frame of image captured after the target object lands for the first time, or the third image and the fourth image are consecutive frames.
[0024] Acquiring the third image specifically includes:
[0025] Sequentially obtain the height of the key point of the target object corresponding to the acquisition moment of each frame of image;
[0026] Calculate the height difference between the key points of two targets corresponding to two adjacent frames of images in sequence;
[0027] When the height difference is smaller than a preset value, the latter frame of two adjacent frames of images corresponding to the height difference is defined as the third image.
[0028] In order to solve the problem of excessive measurement error in the existing technology based on machine vision, the present invention proposes a method for obtaining the first landing point of a flying target, which can reduce the result error. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Figure 1 A schematic diagram of the flight trajectory in a method for acquiring the coordinates of characteristic points of a flying target when the flying target lands for the first time proposed by the present invention. DETAILED DESCRIPTION
[0030] In order to solve the problem that the target object first lands at a large error in the background technology, reference Figure 1 The present invention proposes a method for obtaining the first landing point of a flying target, comprising:
[0031] Multiple frames of images are collected at a fixed point, wherein the multiple frames of images include multiple frames of first type images before the target having the characteristics of the target falls to the ground for the first time and multiple frames of second type images after the target falls to the ground for the first time.
[0032] A first virtual line b1 / b2 is obtained based on multiple frames of first-type images, and the first virtual line b1 / b2 passes through the target feature point when the target first lands within the allowable error range. A second virtual line b3 / b4 is obtained based on one or more frames of second-type images, and the second virtual line b3 / b4 passes through the target feature point when the target first lands within the allowable error range.
[0033] The coordinates of the intersection of the first virtual line b1 / b2 and the second virtual line b3 / b4 are obtained as the coordinates of the characteristic point of the target object when the target object first lands.
[0034] In this specific implementation, regardless of whether the first landing frame is collected, the intersection coordinates a3 / a4 of the first virtual line b1 / b2 and the second virtual line b3 / b4 are used as the first landing coordinates of the target object. Figure 1 , its error is obviously smaller than the calculation method of using a frame in the non-landing frame whose acquisition time is closest to the first landing time of the target, such as a1 or a5, as the landing point to calculate the landing point of the target.
[0035] refer to Figure 1 Furthermore, the first virtual line b1 is a parabola followed by the characteristic points of the target object before the target object first lands, which is fitted based on multiple frames of the first type of images.
[0036] Based on multiple frames of the first type of images, multiple coordinates of the target object before it first lands can be obtained. The target object always roughly follows the parabola rule before it first lands. The first virtual line b1 can be fitted through the above multiple coordinates, and the intersection coordinate a3 of the first virtual line b1 and the second virtual line b3 / b4 is used as the coordinate of the target object's first landing.
[0037] refer to Figure 1 Further, obtaining the first virtual line specifically includes:
[0038] A first image and a second image are acquired from the plurality of frames of first-type images.
[0039] The first coordinate a1 of the key point of the target object at the time of capturing the first image is obtained, and the second coordinate a2 of the key point of the target object at the time of capturing the second image is obtained based on the second image analysis.
[0040] The straight line passing through the first coordinate a1 and the second coordinate a2 is obtained as the first virtual line b2.
[0041] The first image is specifically the last frame of image captured before the target object lands for the first time. Furthermore, the second image and the first image are consecutive frames, that is, the second image is the previous frame of the first image.
[0042] Compared with the existing technology, the accuracy can also be improved.
[0043] Furthermore, acquiring the first image specifically includes:
[0044] The height of the key point of the target object corresponding to the acquisition moment of each frame of image is obtained in sequence.
[0045] The height difference between the key points of two targets corresponding to two adjacent frames of images is calculated in sequence.
[0046] When the height difference is less than a preset value, the previous frame of two adjacent frames of images corresponding to the height difference is defined as the first image.
[0047] It should be pointed out that before the target falls to the ground, gravity is greater than air resistance, and the vertical flight speed gradually increases. Since the time intervals between two adjacent frames are equal, the height difference between two adjacent frames of the target increases gradually. The frames before and after the target falls for the first time drop to a valley value due to touching the ground. Combined with the rebound after touching the ground, the valley value may even be negative. Therefore, when the height difference between adjacent frames is less than the preset value, the previous frame in the corresponding adjacent frame image can be defined as the first image.
[0048] refer to Figure 1 Furthermore, the second virtual line is a parabola followed by the feature points of the target after the target first lands, which is fitted by b4 based on multiple frames of the second type of images.
[0049] Based on multiple frames of the second type of images, multiple coordinates of the target object after it first lands can be obtained. If the target object bounces after it first lands, it can also roughly follow a certain parabolic rule. The second virtual line b4 can be fitted through the above multiple coordinates, and the coordinate a3 of the intersection of the first virtual line b1 / b2 and the second virtual line b4 is used as the coordinate of the target object's first landing.
[0050] refer to Figure 1 Further, the second virtual line b3 is a horizontal straight line, and the second virtual line b3 passes through a target feature point of a target located on the ground.
[0051] Specifically, the coordinates of a target object on the ground can be obtained naturally from multiple frames of the second type of image. The coordinates are not limited to the first landing point of the target object. In practice, the closer the selected point is to the first landing point, the smaller the error. For a solid ball as a target object, its elasticity is not enough to support a second bounce, so it is advisable to Figure 1 The second coordinate a2 in the next frame is the third coordinate a5.
[0052] Further, refer to Figure 1 , obtaining the second virtual line specifically includes:
[0053] A third image and a fourth image are acquired from the multiple frames of the second type of images.
[0054] The third coordinates of the key points of the target object at the time of capturing the third image are obtained, and the fourth coordinates of the key points of the target object at the time of capturing the fourth image are obtained based on the fourth image analysis.
[0055] The straight line passing through the third coordinate and the fourth coordinate is obtained as the second virtual line.
[0056] The third image is specifically the first frame of image collected after the target object lands for the first time. Furthermore, the third image and the fourth image are consecutive frames.
[0057] Compared with the existing technology, the accuracy can also be improved.
[0058] Further, acquiring the third image specifically includes:
[0059] The height of the key point of the target object corresponding to the acquisition moment of each frame of image is obtained in sequence.
[0060] The height difference between the key points of two targets corresponding to two adjacent frames of images is calculated in sequence.
[0061] When the height difference is less than a preset value, the latter frame of two adjacent frames of images corresponding to the height difference is defined as the third image.
[0062] Referring to the description of obtaining the first image, since the previous frame of the two adjacent frames of images corresponding to the height difference is defined as the first image, and the first image is the last frame before the first landing, the latter frame is the first landing frame captured or the first frame after landing if the first landing frame is not captured, both of which are the third image.
[0063] The above are only preferred specific implementation modes of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solutions and inventive concepts of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.
Claims
1. A method for obtaining the coordinates of characteristic points of a flying target when the flying target lands for the first time, characterized in that: include: Collect multiple frames of images at a fixed point, wherein the multiple frames of images include multiple frames of first-type images before the target with target characteristics lands for the first time and one or more frames of second-type images after the target lands for the first time; Acquire a first virtual line according to multiple frames of the first type of image, the first virtual line passing through the target feature point when the target first lands within an allowable error range, and acquire a second virtual line according to one or more frames of the second type of image, the second virtual line passing through the target feature point when the target first lands within an allowable error range; The coordinates of the intersection of the first virtual line and the second virtual line are obtained as the coordinates of the characteristic point of the target object when the target object first lands.
2. The method for obtaining the coordinates of the characteristic points of a flying target when the flying target first lands according to claim 1, characterized in that: The first virtual line is a parabola followed by the feature points of the target object before the target object first lands on the ground, which is fitted based on multiple frames of the first type of images.
3. The method for obtaining the coordinates of the characteristic points of a flying target when the flying target first lands according to claim 1, characterized in that: The step of obtaining the first virtual line specifically includes: Acquire a first image and a second image from the plurality of frames of first type images; Acquire the first coordinates of the key points of the target object at the time of capturing the first image, and analyze the second coordinates of the key points of the target object at the time of capturing the second image according to the second image; The straight line passing through the first coordinate and the second coordinate is obtained as the first virtual line.
4. The method for obtaining the coordinates of the characteristic points of a flying target when the flying target first lands according to claim 3, characterized in that: The first image is specifically the last frame of image captured before the target object lands for the first time, or the second image and the first image are consecutive frames.
5. The method for obtaining the coordinates of the characteristic points of a flying target when the flying target first lands according to claim 4, characterized in that: Acquiring the first image specifically includes: Sequentially obtain the height of the key point of the target object corresponding to the acquisition moment of each frame of image; Calculate the height difference between the key points of two targets corresponding to two adjacent frames of images in sequence; When the height difference is less than a preset value, the previous frame of two adjacent frames of images corresponding to the height difference is defined as the first image.
6. The method for obtaining the coordinates of the characteristic points of a flying target when the flying target lands for the first time according to claim 1, characterized in that: The second virtual line is a parabola followed by the feature points of the target object after the target object first lands, which is fitted based on multiple frames of the second type of images.
7. The method for obtaining the coordinates of the characteristic points of a flying target when the flying target lands for the first time according to claim 1, characterized in that: The second virtual line is a horizontal straight line, and the second virtual line passes through a target feature point of a target located on the ground.
8. The method for obtaining the coordinates of the characteristic points of a flying target when the flying target lands for the first time according to claim 1, characterized in that: The step of obtaining the second virtual line specifically includes: Acquire a third image and a fourth image from the plurality of frames of the second type of images; Acquire the third coordinates of the key points of the target object at the time of capturing the third image, and analyze the fourth coordinates of the key points of the target object at the time of capturing the fourth image according to the fourth image; The straight line passing through the third coordinate and the fourth coordinate is obtained as the second virtual line.
9. The method for acquiring the coordinates of the characteristic points of a flying target when the flying target first lands according to claim 8, characterized in that: The third image is specifically the first frame of image captured after the target object lands for the first time, or the third image and the fourth image are consecutive frames.
10. The method for acquiring the coordinates of the characteristic points of a flying target when the flying target lands for the first time according to claim 9, characterized in that: Acquiring the third image specifically includes: Sequentially obtain the height of the key point of the target object corresponding to the acquisition moment of each frame of image; Calculate the height difference between the key points of two targets corresponding to two adjacent frames of images in sequence; When the height difference is less than a preset value, the latter frame of two adjacent frames of images corresponding to the height difference is defined as the third image.