CN115753610A - Adjusting method for spectrum online detection of placement posture of target object - Google Patents

Abstract

The invention discloses a method for adjusting the posture of a target object in online spectrum detection, which includes creating a target posture detection system and a target posture adjustment system for spectrum online detection; performing calibration of the target posture detection system and the target posture adjustment system ; Acquisition and preprocessing of spectral online detection target source images; feature extraction of preprocessed spectral online detection target images; combined with camera geometry principles to solve spectral online detection target spatial geometric information; according to the solved spectral online detection target The spatial geometric information of the object is compared with the geometric information of the target space, and the target posture adjustment system is used to adjust the spectrum to detect the target posture online. The invention has simple structure, relatively low equipment cost, high detection efficiency, short object attitude adjustment time, can effectively improve the quality detection efficiency of spectrum online detection target object, can well adapt to the needs of large-scale target object quality detection and industrial production, and improves automation Rate.

Description

An adjustment method for online spectrum detection of target object placement posture

technical field

The invention belongs to the technical field of machine vision, and in particular relates to an adjustment method for spectrum online detection of target object placement posture.

Background technique

An important parameter for judging the state of an object in space is attitude. Spectral online target object attitude detection and spectral online target object quality detection after attitude adjustment have a very important technical guidance for industrial automation processing.

Visual inspection technology is a new technology with the most development potential in the field of precision testing technology. It comprehensively uses electronics, photoelectric detection, image processing and computer technology to introduce machine vision into industrial inspection to realize the detection of objects (products or parts) ) The rapid measurement of three-dimensional size or position has outstanding advantages such as non-contact, high speed, and good flexibility, and has important application prospects in modern manufacturing.

Spectrum online detection target attitude adjustment technology enables the object to adjust its posture in a short time, which improves the efficiency and time of spectrum online detection target object quality detection. The attitude adjustment of the existing spectral online detection target is manually adjusted through artificial position calibration, which not only reduces the detection efficiency, but also cannot effectively guarantee the detection quality. At this time, the automatic spectral online detection target posture detection and adjustment method becomes particularly important. The posture correction algorithm is used to realize the automatic spectral online detection target object posture adjustment, so as to meet the spectral online detection target quality detection and industrial automation production.

Contents of the invention

The invention provides an adjustment method for spectrum online detection of target object placement posture. It solves the problem that the attitude adjustment of the existing spectrum online detection target is manually calibrated, and realizes the automatic posture adjustment of the spectrum online detection target through the attitude correction algorithm, increases the degree of automatic detection of the detection device, and improves the efficiency and accuracy of detection.

In order to achieve the above object, the technical solution of the present invention is to provide an adjustment method for spectrum online detection of target object placement posture. The method for online attitude adjustment of spectrum online detection target object comprises the following steps:

Step S1: Create a target posture detection system and a target posture adjustment system for online spectrum detection;

Step S2: Calibrate the target attitude detection system and the target attitude adjustment system;

Step S3: Acquisition and preprocessing of the source image of the target object in the online spectrum detection;

Step S4: feature extraction of the preprocessed spectral online detection target image;

Step S5: Combining the camera geometry principle to solve the spatial geometric information of the spectral online detection target object;

Step S6: Online detection of the spatial geometric information of the target object based on the solved spectrum, comparing the spatial geometric information of the target object, and using the target object posture adjustment system to adjust the spectrum to detect the object posture online.

Preferably, in the step S1, the target posture detection system includes a left CCD camera and a right CCD camera, a light box, a fluorescent lamp, a black background and a computer, and the left CCD camera and the right CCD camera and the fluorescent light are all arranged in the light box, The online detection target of the measured spectrum is placed at the bottom of the light box with a black background.

Preferably, in the step S1, the adjustment object of the target attitude adjustment system is a target tray, the bottom of the tray has a near-infrared light transmission hole, and the outer periphery is provided with 4 rollers, and each roller is connected to a corresponding step. into the motor. The two opposite rollers are respectively connected to the horizontal rotation motor, and the other two opposite rollers are respectively connected to the vertical rotation motor. The horizontal rotation motor and the vertical rotation motor respectively control the horizontal rotation and vertical rotation of the spectrum online detection target to correct the accuracy of the target posture. For stability, the bottom of the tray is parallel to the bottom of the light box.

Preferably, in the step S2, the calibration of the target posture detection system and the target posture adjustment system includes the following steps:

Step S21: Prepare the spectrum for online detection of the target object, use the CCD camera to collect the image of the target object, then perform feature extraction and calculate the spatial geometric parameters;

Step S22: Establish the base coordinate system, and then calculate the parameters of the coordinate system;

Step S23: Obtain the parameters of steps S21 and S22, and convert the spatial geometric parameters of the spectral online detection target obtained in step S21 into the base coordinate system of the target attitude adjustment system. Define the rotation matrix R ^' and translation vector T ^' between the CCD camera and the base coordinates, respectively, where R ^' and T ^' are the initial correction parameters between the camera and the base coordinates. Through the parameter solution, the spectrum under the base coordinate system of the target attitude adjustment system is obtained, and the spatial geometric parameters of the target object are detected online, and compared with the set target attitude space geometric parameters.

Preferably, in the step S3, the preprocessing of the target source image includes image enhancement algorithm, gray scale, and histogram equalization.

Preferably, in the step S4, when performing the feature extraction of the online spectral detection target, the position point, fruit stem or calyx point of the measured spectral online detection target, edge features and size features of the spectral online detection target should be extracted; Among them, the size feature extraction of the target object is extracted using the projected area method.

Preferably, in the step S5, by judging whether there is a fruit stalk or a calyx point, and looking for the coordinates of the centroid point of the target and its symmetry axis to determine the placement method of the target, through the position point of the target image Extract the axis of the image plane with the coordinates of the fruit stem or calyx point, use the space intersection method to solve the three-dimensional attitude of the target axis, use the method of approximately matching the position point to solve the world coordinates of the target position point, and use the spectrum to detect the size of the target object online The formula is used to obtain the actual size of the target object in online detection of the spectrum.

Preferably, in the step S6, when solving the spectral online detection target attitude parameters, there are two cases: when the target is placed vertically, the attitude parameters are directly defined, and there is no need to solve; when the target is placed obliquely, the method of spatial geometry knowledge is used Solve the attitude parameters of the target object.

Preferably, in the step S6, the online spectrum detection target posture parameters are compared with the target posture parameters, if within the error range, the spectrum online detection target posture does not need to be adjusted, otherwise the posture parameter related data is sent to the target object The control mechanism of the attitude adjustment system is adjusted by driving the motor to adjust the rollers of the tray according to the spectral online detection of the target attitude of the target.

As a further supplement to the technical solution, as mentioned above, the adjustment method for spectrum online detection of target placement posture according to the present invention has the following practical beneficial effects:

The method for adjusting the placement attitude of the spectrum online detection target object of the present invention is solved by image acquisition, image preprocessing, feature extraction and the method of spatial intersection, and the projection equation and the spectrum online detection target object through image acquisition processing, feature extraction and spectrum online detection. The plane equation determined by the optical center of the camera is solved to obtain the spatial geometric parameters in the world coordinate system of the online spectral detection target. Using the computer to send data parameters into the execution structure to adjust the attitude of the spectrum online detection target object; the invention has simple structure, relatively low equipment cost, high detection efficiency, short object posture adjustment time, and can effectively improve the quality detection efficiency of the spectrum online detection target object And industrial automation production, can well meet the needs of large-scale target material quality inspection and industrial production.

Description of drawings

Fig. 1 is a flow chart of an embodiment of an online posture adjustment method for an online spectral detection target;

Fig. 2 is a composition diagram of a target attitude detection system;

Fig. 3 is the attitude adjustment target object tray;

Figure 4 is a diagram of the installation position of the roller motor drive;

Fig. 5 is a flowchart of the calibration of the target attitude adjustment system;

Fig. 6 is a flow chart of judgment before target attitude adjustment;

Figure 7 is a data table of the experimental results.

Component label description: 1. Left CCD camera; 2. Right CCD camera; 3. Illumination box; 4. Fluorescent lamp; 5. Black background; 6. Computer; 7. Tray; 8. Near-infrared light transmission hole; 9. Roller installation Position; 10, roller; 11, horizontal rotation motor; 12, vertical rotation motor;

Detailed ways

The implementation mode of the present invention is illustrated by specific examples below, and the advantages and effects of the present invention can be easily understood through the information provided in this description.

It should be noted that the structures, proportions, sizes, etc. drawn in the drawings of this specification are only used to match the content disclosed in the specification, for those who are familiar with this technology to understand and read, and are not used to limit the conditions for the implementation of the present invention. Therefore, there is no technical substantive meaning, and any modification of structure, change of proportional relationship or adjustment of size shall still fall within the scope of the disclosure of the present invention without affecting the functions and objectives of the present invention. within the scope of technical content. At the same time, terms such as "upper", "lower", "left", "right", and "middle" quoted in this specification are only for convenience of description, and are not used to limit the scope of the present invention. , the change or adjustment of its relative relationship, without substantive changes in the technical content, should also be regarded as the scope of the present invention that can be implemented.

The method for adjusting the posture of the target on-line detection of the spectrum of the present invention is characterized in that, firstly, a target posture detection system and a target posture adjustment system of the spectrum online detection are created; secondly, the target posture detection system and the target posture adjustment system established The attitude adjustment system performs system calibration. Use the CCD camera to collect target source images and perform preprocessing. After extracting image features, solve the spectral online detection target spatial geometric information according to the camera geometry principle. After comparing the target spatial geometric information, use the target posture adjustment system to adjust the spectral online detection target posture. , and finally achieve the purpose of attitude correction. The specific process is shown in Figure 1, and specifically includes the following steps:

Step S1: Create a target posture detection system and a target posture adjustment system for online spectrum detection;

Step S2: Calibrate the target attitude detection system and the target attitude adjustment system;

Step S3: Carry out the ingestion and preprocessing of the target source image of the spectrum online detection;

Step S4: feature extraction of the preprocessed spectral online detection target image;

Step S5: Combining the camera geometry principle to solve the spatial geometric information of the spectral online detection target object;

Step S6: Online detection of the spatial geometric information of the target object based on the solved spectrum, comparing the spatial geometric information of the target object, and using the target object posture adjustment system to adjust the spectrum to detect the object posture online.

Specifically, in step S1 , the target posture detection system includes two CCD cameras, a left CCD camera 1 and a right CCD camera 2 , a light box 3 , a fluorescent lamp 4 , a black background 5 , and a computer 6 . In the system, the left CCD camera 1 and the right CCD camera 2 are located above the light box 3, the line connecting the optical centers of the two CCD cameras is parallel to the horizontal side of the light box 3, and the optical axes of the left CCD camera 1 and the right CCD camera 2 are placed at an angle Fluorescent lamps 4 are respectively installed at the four corners of the top of the light box 3, and the target object for online detection of the spectrum to be measured is placed at the bottom of the light box 3 with black as the black background 5. The placement location is shown in Figure 2.

Specifically, in step S1, the adjustment object of the target attitude adjustment system is the target tray 7, the hole at the bottom of the circular tray 7 is a near-infrared light transmission hole 8, and the roller installation position 9 is set around the tray 7, each Roller 10 is connected with corresponding horizontal rotation motor 11 and vertical rotation motor 12 respectively, and horizontal rotation motor 11 and vertical rotation motor 12 respectively control spectral on-line detection object horizontal rotation and vertical rotation; Horizontal rotation motor 11 and vertical rotation motor 12 adopt step Into the motor, the bottom of the tray 7 is parallel to the bottom of the light box 3 and the tray 7 is elevated by four rollers 10, and the corresponding stepping motor is fixed on the table. The attitude adjustment target tray is shown in Figure 3, and the placement position of the stepping motor is shown in Figure 4.

Specifically, in step S2, after the target posture detection system and the target posture adjustment system are created, the target posture detection system needs to be calibrated before capturing images. The system calibration process is shown in FIG. 5 . The left and right CCD camera parameters and system parameters are calibrated, the target attitude adjustment system is calibrated, and the coordinate system parameters and system parameters of the tray 7 and the roller 10 are calibrated.

Specifically, in step S2, the specific steps of the calibration process are:

Step S21: Use the camera to acquire the image of the target object, then perform image feature extraction, and calculate the coordinate parameters;

Two left CCD camera 1 and right CCD camera 2 in the same horizontal plane collect the same calibration image at the same time, respectively calibrate the left CCD camera 1 and right CCD camera 2 parameters and system parameters, and finally give the left CCD camera 1 and right CCD camera 2 in Projection matrices A ₁ and A ₂ in world coordinate system.

The projection matrix of each camera is calibrated according to the principle of space geometry above. The system parameters, that is, the relative positional relationship between the two cameras, also need to be determined. The system parameters are the rotation matrix R and the translation vector T. The rotation and translation relationship between the left CCD camera 1 and the world coordinate system is R _l , T _l , and the rotation and translation relationship between the right CCD camera 2 and the world coordinate system is R _r , T _r .

The coordinates of the world coordinate system are represented by X _w in any point in the space, then the left CCD camera 1 is represented by X _l , and the right phase CCD camera 2 is represented by X _r , then the following equations are established:

According to the mathematical transformation, the rotation matrix R and the translation vector T can be obtained:

Taking the target apple as an example, detect the target image online according to the spectrum collected by the camera, extract the spectrum to detect the feature point of the target on-line, that is, the apple fruit pedicle point or calyx point, and the coordinates of the apple centroid point, and connect these two points with a dotted line Coordinates, the dotted line is the axis of the image plane. Assume that the equation of the image plane axis is ax+by+c=0, and then extract the projection equations of the spectral online detection target on the left and right CCD camera planes respectively as a _l x+b _l y+c _l =0 and a _r x+b _r y+c _r =0. In the respective CCD camera coordinate systems, the focal length of the camera system is f, and the plane equations determined by the projection equation and the camera optical center are respectively:

In the formula, a _l , b _l , c _l represent the coefficients of the camera coordinate equation of the 1-axis of the left CCD camera; a _r , b _r , and c _r represent the coefficients of the camera coordinate equation of the 2-axis of the right CCD camera; x represents the horizontal axis of the feature point Coordinates, y represents the ordinate of the feature point. Convert formula (3) to the equation in the world coordinate system as:

In the formula, A _l , B _l , C _l , D _l represent the coefficients of the world coordinate equation of the left CCD camera 1 axis; A _r , B _r , C _r , D _r represent the coefficients of the world coordinate equation of the right CCD camera 2 axis ;x represents the abscissa of the feature point, y represents the ordinate of the feature point, and z represents the vertical coordinate of the feature point. According to the principle of space geometry, the equation group (4) is the equation of the space axis, and the three directions of the space axis are:

According to the formula (5), the angle between the axis of the online spectrum detection object and the XOY plane projection in the space coordinate system and the angle between the projection of the axis of the spectrum online detection object on the XOY plane and the X axis of space are obtained.

Step S22: Establish the base coordinate system, and then calculate the parameters of the coordinate system;

Step S23: Obtain the parameters of steps S21 and S22, and convert the spatial geometric parameters of the spectral online detection target obtained in step S21 into the base coordinate system of the target attitude adjustment system. Define the rotation matrix R ^' and translation vector T ^' between the CCD camera and the coordinates of the base respectively, and R ^' and T ^' are the initial correction parameters between the coordinates of the camera and the base. Through the parameters, the spectrum in the base coordinate system of the target attitude adjustment system is obtained, and the spatial geometric parameters of the target object are detected online, and compared with the set target attitude space geometric parameters. The comparison detection is shown in Figure 6.

Specifically, in step S23, the comparative detection process is shown in Figure 6, and the specific process is: given the spectrum to detect the target object on-line, judging whether the target object is within the target parameter range, if it is, then end, if not, then the space The parameters are converted into base coordinate parameters, and the system is adjusted for online spectrum detection, and the adjusted spectrum is obtained for online detection of the target object posture, and then it is judged whether the target is within the target parameter range, if it is, it ends, if not, it continues to loop.

Specifically, in step S4, the object of this embodiment is fruit as an example. After the system captures the image of the object for online spectral detection, it preprocesses the image to be suitable for feature extraction of the object for online spectral detection. The preprocessing of the target source image includes image enhancement algorithm, gray scale, and histogram equalization.

Specifically, in step S4, when performing feature extraction of the online spectral detection target, the position point, fruit stem or calyx point of the measured spectral online detection target, edge features and size features of the spectral online detection target should be extracted; wherein, The size feature extraction of the target object is extracted using the projected area method.

Specifically, in step S5, by judging whether there is a fruit stalk or a calyx point, and looking for the coordinates of the centroid point of the target and its symmetry axis on the outer contour of the target, the way to place the target is judged. The coordinates of the stalk or calyx point are used to extract the axis of the image plane, and the spatial intersection method is used to solve the three-dimensional attitude of the target axis. The spectrum detects the actual size of the target object online.

Specifically, in step S6, by comparing the world coordinates of the target position point with the world coordinates of the target coordinates, it is judged whether to adjust the spectrum to detect the target posture online through the target posture adjustment system. There are two situations when solving the attitude parameters of the spectrum online detection target object: when the target is placed vertically, the attitude parameters are directly defined, and there is no need to solve; when the object is placed obliquely, the spatial geometry knowledge method is used to solve the target attitude parameters.

Compare the attitude parameters of the spectrum online detection target with the target attitude parameters. If it is within the error range, the spectrum online detection target attitude does not need to be adjusted. Otherwise, the data related to the attitude parameters will be sent to the control mechanism of the target attitude adjustment system. Adjust the rollers 10 of the circular tray 7 to adjust according to the online detection of the target posture of the spectrum. In this embodiment, after the stepper motor is used to control the rotation of the roller to achieve different rotation angles, the binary image of the fruit contour obtained after the camera captures and processes is shown in FIG. 7 .

The above is only a preferred embodiment of the present invention, it should be pointed out that, for those of ordinary skill in the art, without departing from the principle of the present invention, some improvements and modifications can also be made, and these improvements and modifications can also be made. It should be regarded as the protection scope of the present invention.

Claims (7)

1. An adjusting method for detecting the placing posture of a target object on line through a spectrum. The method is characterized by comprising the following steps:

step S1: establishing a target object posture detection system and a target object posture adjustment system for spectrum online detection;

step S2: calibrating a target object posture detection system and a target object posture adjustment system;

and step S3: collecting and preprocessing a source image of a spectral online detection target object;

and step S4: extracting the characteristics of the preprocessed spectrum on-line detection target object image;

step S5: solving the space geometric information of the spectrum on-line detection target object by combining the shooting geometric principle;

step S6: and detecting the space geometric information of the target object on line according to the solved spectrum, comparing the space geometric information of the target object, and adjusting the spectrum to detect the posture of the target object on line by using a posture adjusting system of the target object.

2. The adjusting method for the spectral online detection of the placement attitude of the target object according to claim 1, characterized in that: in the step S1, the target posture detection system comprises a left CCD camera, a right CCD camera, an illumination box, a fluorescent lamp, a black background and a computer, wherein the left CCD camera, the right CCD camera and the fluorescent lamp are all arranged in the illumination box, and a detected spectrum online detection target is placed at the bottom of the illumination box with the black background.

3. The adjusting method for the spectral on-line detection of the placement posture of the target object according to claim 1, characterized in that: in step S2, the calibration of the target attitude detection system and the target attitude adjustment system includes the following steps:

step S21: preparing a spectrum on-line detection target object, collecting a target object image by using a CCD camera, then performing feature extraction, and calculating space geometric parameters;

step S22: establishing a base coordinate system, and then calculating the parameters of the coordinate system;

step S23: obtaining parameters of S21 and S22, converting the space geometric parameters of the spectrum on-line detection target object obtained in the step S21 into a base coordinate system of a target object posture adjusting system, and respectively defining a rotation matrix R between a CCD camera and the base coordinate ^’ And translation vector T ^’ Wherein R is ^’ And T ^’ Is the initial calibration parameter between the camera and base coordinates. And solving the space geometric parameters of the target object under the base coordinate system of the target object attitude adjusting system through the parameters, and comparing the space geometric parameters with the set space geometric parameters of the target attitude.

4. The adjusting method for the spectral on-line detection of the placement posture of the target object according to claim 1, characterized in that: in step S3, the preprocessing of the target object source image comprises an image enhancement algorithm, graying and histogram equalization.

5. The adjusting method for the spectral on-line detection of the placement posture of the target object according to claim 1, characterized in that: in the step S4, when the characteristic extraction of the spectrum on-line detection target object is carried out, the position point, the fruit stem or the calyx point of the detected spectrum on-line detection target object, the edge characteristic and the size characteristic of the spectrum on-line detection target object are extracted; wherein, the size characteristic of the target object is extracted by using a projection area method.

6. The adjusting method for the spectral online detection of the placement attitude of the target object according to claim 1, characterized in that: in the step S5, the placement mode of the target object is judged by judging whether fruit stems or calyx points exist or not, the coordinate of the center point of the target object and the symmetry axis of the center point are searched for by the outer contour of the target object, the axis of the image plane is extracted through the position points of the image of the target object and the coordinates of the fruit stems or calyx points, the three-dimensional posture of the axis of the target object is solved by adopting a space intersection method, the world coordinate of the position point of the target object is solved by adopting a method of approximately matching the position points, and the actual size of the target object is detected on line by utilizing a spectrum on-line detection target object size solving formula.

7. The adjusting method for the spectral online detection of the placement attitude of the target object according to claim 1, characterized in that: in step S6, solving the spectrum on-line detectionSide surveyWhen the attitude parameters of the target object are measured, two conditions are adopted: when the target is vertically placed, the attitude parameters are directly defined without solving; when the inclined object is placed, the attitude parameters of the object are solved by adopting a method of space geometric knowledge; and comparing the spectrum online detection target object attitude parameter with the target attitude parameter, if the spectrum online detection target object attitude is within the error range, the spectrum online detection target object attitude does not need to be adjusted, otherwise, sending the data related to the attitude parameter into a control mechanism of a target object attitude adjusting system, and adjusting the target object attitude of the target object online detection according to the spectrum by adjusting a roller of a tray through a driving motor.

Images