An improved method for eliminating jitter of road depth images taken with a line laser 3D camera

Abstract

The present invention discloses an improved method for eliminating jitter of a road surface depth image taken by a line laser 3D camera, S1: traverse each line of the depth image; S2: read the depth data and form an array a i , for array a i Sort and get a new array a' i ; S3: array a' i The middle data of the array b i ; S4: Find array b i The average depth data is used to obtain the average ground height dm i , then S5: repeat the above steps until all h rows are traversed, and obtain an array dm consisting of the average ground heights of h rows; S6: low-pass filter dm to filter out the high-frequency signal of dm, and obtain a discrete array dm' containing only the jitter waveform; S7: for each row of depth data in the depth image, subtract the filtered ground height dm' corresponding to the row i , where dm' i is the ith data of the array dm', 1≤i≤h. The beneficial effects of the present invention are: by using an algorithm to eliminate jitter, the operation is greatly simplified, and no additional sensor is required, thereby improving efficiency.

Description

Technical Field

[0001] The invention relates to the field of jitter elimination of depth images, and in particular to an improved method for eliminating jitter of road depth images taken by a line laser 3D camera. Background Art

[0002] A line laser 3D camera is a three-dimensional camera that is based on the principle of triangulation or mechanical light. It captures the laser line information projected by a laser generator on the surface of an object through an image sensor, thereby reconstructing the surface contour information of the object.

[0003] In common factory applications, the laser transmitter and 3D camera are fixed, and the object to be measured passes under the 3D camera at a stable speed on the conveyor belt. Since the 3D camera is fixed, the conveyor belt moves smoothly and the jitter is very small, and its depth measurement accuracy can even reach the micron level. However, in the shooting of the road surface depth map, the image acquisition method has changed. The laser transmitter and 3D camera are fixed on a mobile platform (vehicle or outdoor mobile robot), and follow the moving platform along the photographed road surface. The line laser is perpendicular to the forward direction of the mobile platform. Since the mobile platform inevitably shakes itself during the movement process, the depth image collected by the 3D camera is distorted. Among them, the up and down jitter of the mobile platform perpendicular to the road surface has the greatest impact on the depth image, because the jitter is directly superimposed on the data of the depth image, it will have an adverse effect on automatic disease identification.

[0004] The existing method is to add an accelerometer (or IMU, inertial navigation, etc., which are essentially acceleration measurements) to the 3D acquisition system, measure the time-distributed acceleration of the entire 3D acquisition system in the direction perpendicular to the road surface through the accelerometer, and then perform a second integration of the acceleration to obtain the time-distributed displacement in the direction perpendicular to the road surface (i.e., jitter interference). Each row on the acquired depth image also corresponds to a specific time. Finally, the acquired depth image is subtracted from the corresponding displacement after interpolation to filter out the jitter. However, this method has certain defects. First, additional hardware needs to be installed, and the structural rigidity needs to be ensured so that the jitter of the sensor and the 3D camera are completely consistent; secondly, the acceleration needs to be integrated twice to obtain the displacement change, and the cumulative error is large, which affects the accuracy of the calculated jitter displacement. After long-term research by the inventor, an improved method for eliminating jitter in road depth images taken with a line laser 3D camera has been invented. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and provide an improved method for eliminating jitter of road depth images taken with a line laser 3D camera.

[0006] The purpose of the present invention is achieved by the following technical solutions: an improved method for eliminating jitter of a road depth image taken by a line laser 3D camera, assuming that the road depth image taken by the 3D camera has h rows and w columns, and the depth data corresponding to each pixel is d ij , where i represents the i-th row, j represents the j-th column, 1≤i≤h, 1≤j≤w, including the following steps:

[0007] S1: traverse each row of the depth image, assuming that the current row is the i-th row;

[0008] S2: Read all w depth data of row i and form an array a i , for array a i Sort the depth data from small to large to get a new array a' i ;

[0009] S3: intercept array a' i The middle data of the array is used to form a new array b i ;

[0010] S4: Find array b i The average ground height corresponding to the i-th row is obtained by averaging the depth data. Let the average ground height of this row be dm i ,but

[0011]

[0012] where b in is array b i The depth data corresponding to the nth element;

[0013] S5: Repeat the above steps until all h rows are traversed, and obtain an array dm consisting of the average ground heights of h rows;

[0014] S6: low-pass filter dm to filter out the high-frequency signal of dm, and obtain a discrete array dm' containing only the jitter waveform;

[0015] S7: For each row of depth data in the depth image, subtract the filtered ground height dm' corresponding to the row i , where dm' i It is the i-th data of the array dm', 1≤i≤h, that is, the new depth image after jitter elimination is obtained.

[0016] Preferably, the low-pass filter cutoff frequency in S7 is higher than the depth picture jitter frequency.

[0017] Preferably, the method for obtaining the middle data of S3 comprises the following steps:

[0018] A1: Assume that k data are taken from the middle segment, where 1<k<w;

[0019] A2: Order

[0020]

[0021] Then the array a' i Remove the first x data and the last y data, and the remaining k data can form b i .

[0022] Preferably, n≤k and n≥1 in S4.

[0023] The present invention has the following advantages: the present invention eliminates jitter by utilizing an algorithm, thereby greatly simplifying the calculation, and does not require the installation of additional sensors, thereby improving efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] none. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0026] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0028] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0029] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inside", "outside", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, or the positions or positional relationships in which the product of the invention is usually placed when in use, or the positions or positional relationships commonly understood by those skilled in the art, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0030] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0031] In this embodiment, an improved method for eliminating jitter of a road depth image taken by a line laser 3D camera is provided. Assume that the road depth image taken by the 3D camera has h rows and w columns, and the depth data corresponding to each pixel is d ij , where i represents the i-th row, j represents the j-th column, 1≤i≤h, 1≤j≤w, including the following steps:

[0032] S1: traverse each row of the depth image, assuming that the current row is the i-th row;

[0033] S2: Read all w depth data of row i and form an array a i , for array a i Sort the depth data from small to large to get a new array a' i ;

[0034] S3: intercept array a' i The middle data of the array is used to form a new array b i ;

[0035] S4: Find array b i The average ground height corresponding to the i-th row is obtained by averaging the depth data. Let the average ground height of this row be dm i ,but

[0036]

[0037] where bin is array b i The depth data corresponding to the nth element;

[0038] S5: Repeat the above steps until all h rows are traversed, and obtain an array dm consisting of the average ground heights of h rows;

[0039] S6: low-pass filter dm to filter out the high-frequency signal of dm, and obtain a discrete array dm' containing only the jitter waveform;

[0040] S7: For each row of depth data in the depth image, subtract the filtered ground height dm' corresponding to the row i , where dm' i is the i-th data of the array dm', 1≤i≤h, that is, the new depth image after the jitter is eliminated is obtained. By using the algorithm to eliminate the jitter, the operation is greatly simplified, and there is no need to install sensors, which improves the efficiency. Specifically, according to the shooting frequency of the 3D camera, dm can be regarded as a discrete waveform of the ground height changing with time, and the low-pass filtering of dm in step S7 is because when there are grooves, roughness and plate seams on the road surface, the grooves, roughness and plate seams can reflect that the depth data in the direction of movement of the mobile platform are all local mutations, which belong to high-frequency signals, while the jitter of the mobile platform itself belongs to a slowly changing low-frequency signal. Therefore, the high-frequency signal of dm can be filtered out by low-pass filtering, thereby ensuring accuracy.

[0041] Furthermore, the low-pass filter cutoff frequency in S7 is higher than the depth picture jitter frequency.

[0042] In this embodiment, the method for obtaining the middle data of S3 includes the following steps:

[0043] A1: Assume that k data are taken from the middle segment, where 1<k<w, for example, k can be w / 2;

[0044] A2: Order

[0045]

[0046] Then the array a' i Remove the first x data and the last y data, and the remaining k data can form b i Specifically, the purpose of taking the middle data is to eliminate the maximum and minimum values ​​so that the data can better reflect the actual ground height.

[0047] Furthermore, in S4, n≤k and n≥1.

[0048] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An improved method for eliminating jitter of road depth images taken by a line laser 3D camera, characterized in that: Assume that the road depth image taken by the 3D camera has h rows and w columns, and the depth data corresponding to each pixel is d ij , where i represents the i-th row, j represents the j-th column, 1≤i≤h, 1≤j≤w, including the following steps: S1: traverse each row of the depth image, assuming that the current row is the i-th row; S2: Read all w depth data of row i and form an array a i , for array a i Sort the depth data from small to large to get a new array a' i ; S3: intercept array a' i The middle data of the array is used to form a new array b i ; S4: Find array b i The average ground height corresponding to the i-th row is obtained by averaging the depth data. Let the average ground height of this row be dm i ,but where b in is array b i The depth data corresponding to the nth element; S5: Repeat the above steps until all h rows are traversed, and obtain an array dm consisting of the average ground heights of h rows; S6: low-pass filter dm to filter out the high-frequency signal of dm, and obtain a discrete array dm' containing only the jitter waveform; S7: For each row of depth data in the depth image, subtract the filtered ground height dm' corresponding to the row i , where dm' i It is the i-th data of the array dm', 1≤i≤h, that is, the new depth image after jitter elimination is obtained.

2. According to claim 1, the improved method for eliminating jitter of road depth images taken by a line laser 3D camera is characterized in that: The low-pass filter cutoff frequency in S7 is higher than the depth picture jitter frequency.

3. The improved method for eliminating jitter of a road depth image taken by a line laser 3D camera according to any one of claims 1 to 2, characterized in that: The method for obtaining the middle data of S3 comprises the following steps: A1: Assume that k data are taken from the middle segment, where 1<k<w; A2: Order y=wkx, then array a' i Remove the first x data and the last y data, and the remaining k data can form b i .

4. The improved method for eliminating jitter of road depth images taken by a line laser 3D camera according to claim 3 is characterized in that: In S4, n≤k, and n≥1.

Images