Laser-fused dual-wavelength all-pose visual sensor

Abstract

The invention discloses a laser-fused dual-wavelength all-pose visual sensor and a method. The sensor adopts the camera, the lens and the LED light source, and can realize the measurement of up-down and left-right translation and rotation angle of a target through the reflecting plate. Meanwhile, four lasers parallel to the optical axis of the parallel lens are adopted to form upper, lower, left and right laser triangulation ranging units, and measurement of the target distance, up-and-down pitching and left-and-right swing angles can be achieved. Therefore, non-contact high-precision detection of the three-dimensional position and the three-dimensional attitude of the target can be realized. In addition, the wavelengths of the laser and the LED light source are red light and blue light respectively, the camera is a color camera, laser triangulation ranging and monocular vision measurement can be achieved at the same time through RGB components of one image, and therefore 6D all-pose synchronous detection is achieved.

Description

Technical Field

[0001] The present invention relates to a posture sensor and method, in particular to a dual-wavelength full-posture visual sensor integrating laser. Background Art

[0002] Target position and posture detection is one of the common detection needs in many fields of the national economy, such as space mechanism docking, robot end positioning, robot random grasping, intelligent press assembly, etc.

[0003] Conventional position and attitude detection are carried out separately, that is, the target three-dimensional position and three-dimensional attitude are detected separately. There are many main methods for three-dimensional position detection, such as laser triangulation ranging sensors, laser trackers, three-coordinate measuring machines, total stations, etc. There are fewer methods for three-dimensional attitude detection, such as photoelectric collimation method, multi-point reflector method, etc. Some scholars in Jinnan have proposed an attitude detection method based on deep learning, but it is still in the laboratory research stage. Especially low, there is a lack of means and methods for 6D full posture parameters that can directly detect the three-dimensional position and three-dimensional attitude of the target at home and abroad. Therefore, there is an urgent need for a high-precision, integrated 6D full posture parameter detection method to meet the application needs of multiple scenarios. Summary of the invention

[0004] In view of the current situation of lack of 6D posture detection means, the present invention proposes a dual-wavelength full-posture visual sensor integrating laser. The visual sensor adopts a camera, a lens and an LED light source, and can realize the measurement of three degrees of freedom of the target's up and down, left and right translation and rotation angle through a reflector. At the same time, four lasers parallel to the optical axes of the camera and the lens are used to form four laser triangulation units, which can realize the measurement of three degrees of freedom of target distance, up and down pitch and left and right swing angle. In this way, non-contact high-precision detection of a total of six degrees of freedom of the target's three-dimensional position and three-dimensional posture can be achieved. In addition, the wavelengths of the laser and the LED light source are red light and blue light respectively, and the camera is a color camera. Laser triangulation and monocular vision measurement can be realized simultaneously through the RGB components of an image, thereby realizing the synchronous detection of 6D full posture.

[0005] The present invention is achieved through the following technical solutions: The special feature of the dual-wavelength omni-position vision sensor of the present invention is that the sensor is composed of a vision probe and a reflector, and the vision probe includes a camera, a lens, a laser, an LED light source, a filter, a processing circuit and a housing, wherein: The camera is an industrial-grade camera, which is responsible for acquiring the image of the reflector; the camera is a color camera, which can extract RGB components; The lens is an industrial-grade high-resolution lens, which is responsible for imaging the image of the reflector onto the image plane of the camera; The laser is an industrial-grade semiconductor laser, and there are four of them, which are located around the lens, arranged in the up, down, left, and right directions, and spaced 90 degrees apart. The outgoing laser beams of the four lasers are parallel to the optical axis of the camera and the lens; the wavelength of the laser is red light or blue light; The LED light sources are multiple and evenly arranged around the lens and spaced apart from the laser; the central wavelength of the LED light source is red light or blue light, but must be different from the wavelength of the laser; The filter is a dual-band narrow-band filter, the central wavelengths of its two passbands are respectively consistent with the central wavelength of the laser and the central wavelength of the anti-LED light source, and the bandwidth of the two passbands is less than 1 / 4 of the difference between the two wavelengths; the filter is located at the front end of the visual probe and is used to filter out other environmental interference light; The processing circuit is a measurement and control circuit based on a high-performance microprocessor, which is responsible for controlling the laser and LED light source on the one hand, and processing the color image data of the camera on the other hand, extracting the R component and the B component, and obtaining the required 6 posture parameters at the same time; The shell is made of a highly conductive metal material, used to support and fix all internal components, and to provide reliable protection and shielding. The reflector is a three-layer structure: the middle layer is a substrate with a diffuse reflection surface, which ensures that the laser beams emitted by four lasers can be reliably irradiated on the substrate; the upper layer is a reflective mark made of retroreflective material and smaller in size than the substrate, including a central reflective mark with a symmetrical shape located in the center, and four straight reflective marks evenly distributed around the central reflective mark at 90°; the lower layer is a strong magnetic plate, one side of which is bonded to the diffuse reflection substrate, and the other side is magnetically fixed to a metal plane of the target to be measured, to ensure that the reflector moves with the target to be measured.

[0006] The working process of the dual-wavelength full-pose visual sensor of the fusion laser of the present invention is as follows: First, the reflector is adsorbed on a certain surface of the target to be measured. Then, under the control of the processing circuit, four LED light sources are turned on to illuminate the reflector, and four lasers are turned on at the same time to generate four light spots on the surface of the reflector. The camera then acquires the image of the reflector area and sends it to the processing circuit for processing. The G component in the color image is eliminated, and the R component and the B component are retained. The R component and the B component of the color image signal are processed separately, and the up and down, left and right translation and angle information corresponding to the monocular vision and the distance information, up and down pitch information, and left and right swing angle information corresponding to the laser triangulation can be obtained synchronously, a total of 6 degrees of freedom parameters, so that the position and attitude signals can be obtained.

[0007] The special feature of the dual-wavelength omni-position visual sensor of the fusion laser of the present invention is that the reflective mark in the reflective plate can adopt a variety of combined symmetrical shapes, or form an integrated reflective mark.

[0008] The present invention proposes a data processing method for the dual-wavelength full-pose vision sensor for the above-mentioned fusion laser, which can simultaneously obtain the coordinate values ​​of three directions and the angle values ​​of three rotation axes, a total of six degrees of freedom measurement results, as follows: (1) The processing circuit turns on four LED light sources to illuminate the reflector, and turns on four lasers to form four laser spots on the surface of the reflector; (2) The camera captures an image that contains both the red or blue laser spot and the reflection mark; (3) Eliminate the G component in the color image and retain the R and B components where the laser spot and reflection mark are located; (4) Processing the R component (or B component) of the color image to obtain the coordinate values ​​of the center points of four laser spots, and using the pre-stored calibration curve to obtain four distance values; (5) Take the average of the above four distance values ​​as the distance value between the visual probe and the reflector; (6) Take the difference between the upper and lower distance values ​​of the above four distance values, divide it by the distance between the two laser spots, and then calculate the arc tangent to obtain the upper and lower pitch angle value of the reflector; (7) Take the difference between the left and right distance values ​​of the above four distance values, divide it by the distance between the two laser spots, and then calculate the arc tangent to obtain the left and right swing angle value of the reflector; (8) Processing the B component (or R component) of the color image to obtain the coordinates of the circular reflective mark, and using the pre-stored measurement curve to obtain the coordinate values ​​of the reflector in the up and down and left and right directions; (9) Process the B component (or R component) of the color image to obtain the angles between the four straight reflection marks and the upper, lower, left and right coordinate axes, and take the average of the two angles as the rotation angle value of the reflection plate. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 It is a schematic diagram of the composition of the dual-wavelength full-pose visual sensor of the fusion laser of the present invention; Figure 2 It is a schematic diagram of the layout of the laser and LED light source of the full-pose visual sensor of the present invention; Figure 3 is a schematic diagram of the passband of the optical filter of the present invention; Figure 4 It is a schematic diagram of the composition of the reflector of the present invention; Figure 5 A schematic diagram of an image captured by the camera of the present invention; Figure 6 A schematic diagram of an image captured by the camera of the present invention; Figure 7 It is a schematic diagram of the distance value, pitch angle and yaw angle measurement principle of the present invention; Figure 8 It is a schematic diagram of the translation coordinate and rotation angle measurement principle of the present invention; In the figure, 1 is a visual probe, 2 is a reflector, 3 is a camera, 4 is a lens, 5 is a laser, 6 is an LED light source, 7 is a filter, 8 is a processing circuit, 9 is a housing, 10 is a substrate, 11 is a reflective mark, and 12 is a magnetic plate. DETAILED DESCRIPTION

[0010] The following is a detailed description of an embodiment of the present invention in conjunction with the accompanying drawings. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and a specific operation process, but the protection scope of the present invention is not limited to the following embodiment.

[0011] The special feature of the dual-wavelength omni-position vision sensor of the present invention is that the sensor is composed of a vision probe 1 and a reflector 2, and the vision probe 1 includes a camera 3, a lens 4, a laser 5, an LED light source 6, a filter 7, a processing circuit 8 and a housing 9. Figure 1 As shown, where: The camera 3 is an industrial-grade camera responsible for acquiring the image of the reflector 2; the camera 1 is a color camera that can extract RGB components; The lens 4 is an industrial-grade high-resolution lens, which is responsible for imaging the image of the reflector 2 onto the image plane of the camera 3; The laser 5 is an industrial-grade semiconductor laser, and there are four of them, namely 5-1, 5-2, 5-3, and 5-4, which are respectively located around the lens 4, in the up and down, left and right directions, and arranged at intervals of 90°. The outgoing laser beams of the four lasers 5-1, 5-2, 5-3, and 5-4 are parallel to the optical axes of the camera 3 and the lens 4; the wavelength of the laser 4 is red light or blue light, for example, 650nm red light can be selected; The LED light sources 6 are multiple and are evenly arranged around the lens 4 and spaced apart from the laser 5. Figure 2 As shown, there are four LED light sources 6-1, 6-2, 6-3, and 6-4, which are arranged between the four lasers 5-1, 5-2, 5-3, and 5-4; the central wavelength of the LED light source 6 is red light or blue light, but must be different from the wavelength of the laser 5; for example, if the laser wavelength is 650nm red light, the LED light source can use 405nm blue-violet light; The filter 7 is a dual-band narrowband filter, the center wavelengths of its two passbands should be consistent with the center wavelengths of the laser 5 and the LED light source 6, respectively, and the bandwidths of its two passbands are less than 1 / 4 of the difference between the two wavelengths. Figure 3 As shown; for example, for the laser 5 with a wavelength of 650nm and the LED light source 6 with a wavelength of 405nm, the center wavelengths of the two passbands of the filter 7 are 650nm and 405nm respectively, and the bandwidth does not exceed 60nm; the filter 7 is located at the front end of the sensor to filter out other interfering light; The processing circuit 8 is a circuit based on a high-performance microprocessor, which is responsible for controlling the laser 5 and the LED light source 6 on the one hand, and for processing the image data of the camera 3 on the other hand, extracting the R component and the B component, and finally obtaining the required 6 posture parameters; The housing 9 is made of a highly conductive metal material, such as aluminum alloy, stainless steel, copper, etc., and is used to support and fix all internal components and provide reliable protection and shielding. The reflector 2 has a three-layer structure. Figure 4 As shown: the middle layer is a substrate 10 with a diffuse reflection surface, which ensures that the laser beams emitted by the four lasers 5-1, 5-2, 5-3, and 5-4 can be reliably irradiated on the substrate 10; the upper layer is a reflective mark 11 made of retroreflective material and smaller in size than the substrate 10, including a central reflective mark 11-1 located in the center and having a symmetrical shape, and four straight reflective marks 11-2 evenly distributed around the central reflective mark at 90°; the lower layer is a strong magnetic plate 12, one side of which is bonded to the diffuse reflection substrate 10, and the other side is magnetically fixed to a metal plane of the target to be measured, so as to ensure that the reflective plate 2 moves with the target to be measured.

[0012] The working process of the dual-wavelength full-attitude visual sensor of the fusion laser of the present invention is as follows: under the control of the processing circuit 8, first turn on the four LED light sources 6 to illuminate the reflector 2, and at the same time turn on the four lasers 5 to generate four laser spots on the surface of the reflector 2. Then the camera 3 obtains the image of the area where the reflector 2 is located, and sends it to the processing circuit 8 for processing. The G component in the color image is eliminated, and the R component and the B component are retained. The R component and the B component of the color image signal are processed separately, and the up and down, left and right translation and angle information corresponding to the monocular vision and the distance information, up and down pitch information, and left and right swing angle information corresponding to the laser triangulation can be obtained synchronously, so that the position and attitude signals can be obtained, thereby realizing zero position calibration.

[0013] The special feature of the dual-wavelength omni-position visual sensor integrating laser of the present invention is that the reflective mark 11 in the reflective plate 2 can adopt a variety of combined symmetrical shapes, or can form an integrated reflective mark 11.

[0014] The present invention proposes a data processing method for the dual-wavelength full-pose vision sensor for the above-mentioned fusion laser, which can simultaneously obtain the coordinate values ​​of three directions and the angle values ​​of three rotation axes, a total of six degrees of freedom measurement results, as follows: (1) The processing circuit 8 turns on the four LED light sources 6-1, 6-2, 6-3, and 6-4 to illuminate the reflector 2, and at the same time turns on the four lasers 5-1, 5-2, 5-3, and 5-5 to form four laser spots on the surface of the reflector 2d; (2) Camera 3 captures an image, such as Figure 5 As shown, the image includes red or blue laser spots 13-1, 13-2, 13-3, 13-4 and reflective marks 11-1, 11-2; (3) Eliminate the G component in the color image, and retain the R component and B component where the laser spot 13 and the reflective mark 11 are located; for example, if the wavelength of the laser 5 is red light, the laser spot 13 is in the R component; if the wavelength of the LED spot is blue light, the reflective mark 11 is in the B component; (4) Process the R component of the color image to obtain the center point coordinates L1(x1,y1), L2(x2,y2), L3(x3,y3), L4(x4,y4) of the four laser spots 13-1, 13-2, 13-3, and 13-4, and use the pre-stored calibration curve to obtain four distance values ​​Z1, Z2, Z3, and Z4, as shown in Figure 6 As shown; (5) Take the average value of the above four distance values ​​Z1, Z2, Z3, and Z4 as the distance value Z between the visual probe 1 and the reflector 2, that is, Z=(Z1+Z2+Z3+Z4) / 4; (6) Take the difference between the upper and lower distance values ​​Z1 and Z3 of the above four distance values, divide it by the distance D between the two laser spots 13-1 and 13-3, and then calculate the arc tangent to obtain the upper and lower pitch angle values ​​β of the reflector 2, that is, β = arctan[(Z1–Z3) / D]; (7) Take the difference between the left and right distance values ​​Z2 and Z4 of the above four distance values, divide it by the distance D between the two laser spots 13-2 and 13-4, and then calculate the arc tangent to obtain the left and right deflection angle value γ of the reflector, that is, γ = arctan[(Z2 –Z4) / D]; (8) Process the color image B component to obtain the coordinates C (xc, yc) of the circular reflective mark 11-1, and use the pre-stored measurement curve to obtain the coordinate values ​​Xc and Yc of the reflector 2 in the up and down and left and right directions, as shown in FIG. Figure 7 As shown; (9) The color image B component is processed to obtain the angles α1 and α2 between the four linear reflection marks 11 - 1 and the upper, lower, left, and right coordinate axes XY. The average value of the two angles α1 and α2 is taken as the rotation angle value α of the reflection plate 2, that is, α=(α1+α2) / 2.

[0015] Compared with the prior art, the present invention has the following outstanding technical advantages: (1) Since the omni-orientation sensor of the present invention adopts a four-beam parallel laser triangulation ranging method to obtain the required distance information, and adopts monocular vision and reflection marks to obtain translation position and rotation angle information, the three coordinate values ​​and three rotation angle values ​​of the reflection plate are obtained by calculation, with a total of six degrees of freedom results. It not only has many detection parameters and high measurement accuracy, but also has no effect on the target being measured, and has the best versatility.

[0016] (2) Since the full-position and posture sensor of the present invention adopts dual wavelengths to realize multi-detection of six parameters respectively, all six parameter information can be synchronously obtained in one image, which is not only fast but also synchronously processed, which is more reasonable.

Claims

1. A dual-wavelength omni-position vision sensor fused with laser is special in that the sensor is composed of a vision probe and a reflector, and the vision probe includes a camera, a lens, a laser, an LED light source, a filter, a processing circuit and a housing, wherein: The camera is an industrial-grade camera, which is responsible for acquiring the image of the reflector; the camera is a color camera, which can extract RGB components; The lens is an industrial-grade high-resolution lens, which is responsible for imaging the image of the reflector onto the image plane of the camera; The laser is an industrial-grade semiconductor laser, and there are four of them, which are located around the lens, arranged in the up, down, left, and right directions, and spaced 90 degrees apart. The outgoing laser beams of the four lasers are parallel to the optical axis of the camera and the lens; The wavelength of the laser is red light or blue light; The LED light sources are multiple and evenly arranged around the lens and spaced apart from the laser; the central wavelength of the LED light source is red light or blue light, but must be different from the wavelength of the laser; The filter is a dual-band narrow-band filter, the central wavelengths of its two passbands should be consistent with the central wavelength of the laser and the central wavelength of the anti-LED light source respectively, and the bandwidth of the two passbands is less than 1 / 4 of the difference between the two wavelengths; the filter is located at the front end of the sensor to filter out other interfering light; The processing circuit is a circuit based on a high-performance microprocessor, which is responsible for controlling the laser and LED light source on the one hand, and for processing the color image data of the camera on the other hand, extracting the R component and the B component, and obtaining the required 6 posture parameters at the same time; The shell is made of a highly conductive metal material, used to support and fix all internal components, and to provide reliable protection and shielding. The reflector is a three-layer structure: the middle layer is a substrate with a diffuse reflection surface, which ensures that the laser beams emitted by four lasers can be reliably irradiated on the substrate; the upper layer is a reflective mark made of retroreflective material and smaller in size than the substrate, including a central reflective mark with a symmetrical shape located in the center, and four straight reflective marks evenly distributed around the central reflective mark at 90°; the lower layer is a strong magnetic plate, one side of which is bonded to the diffuse reflection substrate, and the other side is magnetically fixed to a metal plane of the target to be measured, to ensure that the reflector moves with the target to be measured.

2. The special feature of the dual-wavelength omni-position visual sensor with fusion laser according to claim 1 is that the reflective mark in the reflective plate can adopt a variety of combined symmetrical shapes, or form an integrated reflective mark.

3. The present invention proposes a data processing method for the dual-wavelength full-position vision sensor of the above-mentioned fusion laser, which can simultaneously obtain the coordinate values ​​of three directions and the angle values ​​of three rotation axes, a total of six degrees of freedom measurement results, as follows: (1) The processing circuit turns on four LED light sources to illuminate the reflector, and turns on four lasers to form four laser spots on the surface of the reflector; (2) The camera captures an image that contains both the red or blue laser spot and the reflection mark; (3) Eliminate the G component in the color image and retain the R and B components where the laser spot and reflection mark are located; (4) Processing the R component (or B component) of the color image to obtain the coordinate values ​​of the center points of four laser spots, and using the pre-stored calibration curve to obtain four distance values; (5) Take the average of the above four distance values ​​as the distance value between the sensor and the reflector; (6) Take the difference between the upper and lower distance values ​​of the above four distance values, divide it by the distance between the two laser spots, and then calculate the arc tangent to obtain the upper and lower pitch angle value of the reflector; (7) Take the difference between the left and right distance values ​​of the above four distance values, divide it by the distance between the two laser spots, and then calculate the arc tangent to obtain the left and right swing angle value of the reflector; (8) Processing the B component (or R component) of the color image to obtain the coordinates of the circular reflective mark, and using the pre-stored measurement curve to obtain the coordinate values ​​of the reflector in the up and down and left and right directions; (9) Process the B component (or R component) of the color image to obtain the angles between the four straight reflection marks and the upper, lower, left and right coordinate axes, and take the average of the two angles as the rotation angle value of the reflection plate.

Images