CN107179317B

CN107179317B - Self-adaptive method for image threshold of retainer suitable for judging front and back surfaces of retainer

Info

Publication number: CN107179317B
Application number: CN201710260504.0A
Authority: CN
Inventors: 谢毅
Original assignee: Zhejiang Gongshang University
Current assignee: Zhejiang Gongshang University
Priority date: 2017-04-20
Filing date: 2017-04-20
Publication date: 2020-05-26
Anticipated expiration: 2037-04-20
Also published as: CN107179317A

Abstract

The invention discloses a self-adaptive method of a holder image threshold value suitable for judging the front and back surfaces of a holder, which is characterized in that the brightness of an end surface recognition LED lamp is not changed when the holder is judged, a self-adaptive threshold value method is adopted, the threshold value of an image shot by a camera is set to change within an allowable range, the change range of the threshold value is defined as a standard threshold value +/-7, the standard threshold value is defined as the most appropriate threshold value of a holder standard part determined during self-learning and is marked as T, the allowable adjustment range of the threshold value is 15 values in total, namely T-7, …, T, … and T +7, when the holder image is judged, the end surface of the holder is shot firstly, the threshold value is removed as T, then the image is judged, the invention adopts the self-adaptive threshold value method, the threshold value of the image shot by the camera is set to change within the allowable range, the adjustment is only needed for the end, the adjustment of 15 brightness levels can be realized, and the image judgment can be realized in the shortest time.

Description

Self-adaptive method for image threshold of retainer suitable for judging front and back surfaces of retainer

Technical Field

The invention relates to the technical field of mechanical part detection equipment, in particular to a self-adaptive method for a retainer image threshold value, which is suitable for judging the front side and the back side of a retainer.

Background

The cage partially wraps all or part of the rolling bodies and moves therewith to isolate the rolling bodies, guide them and retain them within the bearing. For cages for devices with high rotational speeds, such as automobile transmissions, the demands on dimensional accuracy are high. If the size precision of the retainer does not reach the standard in the production process, the retainer is arranged on a vehicle, so that transitional abrasion of the retainer is caused, the service life of the ball cage is influenced, or the running safety and the like are caused. The retainer is a metal component with spherical inner and outer surfaces and two flat ends, which is formed by turning, in the existing retainer processing technology, the front and back surfaces of the retainer are generally judged by manual mechanical measurement, and then the retainer can enter processing equipment in the same direction when the next processing procedure is carried out. In the practice of mass production, on one hand, due to the large workload and the high labor intensity, the condition of missed judgment or misjudgment may occur in manual judgment, the difference between the front side and the back side of the cut retainer is not obvious, and if the precision of the measuring tool is not high enough, the front side and the back side of the retainer cannot be accurately judged, that is, the accurate judgment rate of the front side and the back side of each retainer cannot be ensured.

China with the patent application number of CN2015102506173 discloses a device for judging the front and back surfaces of a retainer, the width of a chamfer of a central circular hole of the retainer is shot through a camera, when the central axis of the camera forms an included angle of 45 degrees with the horizontal plane, the chamfer surface of the central circular hole of the retainer is just vertical to the camera, the chamfer width shot by the camera is the actual width, the front and back surfaces of the retainer can be judged according to the comparison between the chamfer width shot at the moment and the standard chamfer width, the judgment efficiency is high, and the problems of large manual judgment error and low judgment efficiency are solved. However, in practical production tests, we have found that the scheme has the following problems:

1) because the front and back sides of the retainer are characterized by the size of the inner chamfer, the chamfer image shooting must be realized by a 45-degree light source to realize image processing; however, the requirement of chamfer shooting on a light source is high, and for some holders with smaller chamfers, the shot image ring is very narrow, so that misjudgment is easy to occur;

2) in the production process and the automatic feeding process of the retainer, a lot of oil stains are placed on the surface of the retainer and at the chamfer of the retainer, so that the captured image is unclear, and the ring image is broken seriously and even so that the image recognition is wrong;

3) the equipment is required to have lower cost when being applied to a production link, the retainer is grabbed and placed below the camera through a mechanical claw formed by the cylinder and the gas claw, and meanwhile, a manipulator can adapt to the retainers with different diameters as much as possible, so that the positioning precision of the retainer is easy to cause low, and the situation that the light source reflection intensity is unstable due to the fact that the angle deflection possibly occurs after the retainer is grabbed, thereby affecting the quality of images;

4) the retainer can cause the circle centers of the outer circle and the inner circle to be not completely overlapped due to the processing reason, the method of only taking one section of the circle for identification cannot be used, and the whole image must be identified; in consideration of the cost factor of the camera, it is impossible to adopt a camera having a high resolution in an actual production process, thereby also causing difficulty in image recognition.

Meanwhile, the whole system has very high requirements on the reliability of image recognition, the situation of misjudgment in the production process is absolutely not allowed, and the number of suspicious products must be reduced, so that the equipment cannot be well adapted to the actual production process. The applicant designs a method for simultaneously shooting pictures of the end face and the chamfer of the retainer and distinguishing the front face and the back face of the retainer by using AND operation, but in an actual test, the purchased LED lamp is found to work in a transverse flow mode, the brightness of light changes when the current changes, the temperature of the LED lamp changes along with the prolonging of the service time of the LED lamp, and the brightness of the LED lamp changes, so that the image changes, therefore, errors are easily generated in the image binarization processing process to cause image feature extraction errors, for example, the width of an image-identified ring is large when the brightness of the LED lamp is strong, other rings are small or the surface of the ring is broken when the brightness of the LED lamp is weak, and the complete whole ring cannot be identified. The conventional method is a method for automatically determining a threshold value through a histogram, but the method automatically identifies the edge of an image in practical use, but the identified edge is not necessarily the actually required edge, so that the edge is determined without a characteristic inherent to a workpiece as a basis and is easily interfered by other pixels of the image.

Disclosure of Invention

The invention aims to solve the problems and provides a holder image threshold value self-adaption method suitable for holder front and back surface discrimination, which can avoid misjudgment, prevent other items from interference and improve the accuracy of holder front and back surface discrimination

The invention realizes the purpose through the following technical scheme: a self-adaptive method for a holder image threshold value suitable for distinguishing the front side and the back side of a holder is characterized in that the front side and the back side of the holder are distinguished by adopting an end face recognition LED lamp for illumination, a camera is used for shooting, the camera is arranged right above the end face recognition LED lamp, and the axial lead of the camera and the axial lead of the end face recognition LED lamp are on the same straight line; the end face recognition LED lamp comprises an annular lamp cover, a ring of annular point light sources arranged inside the lamp cover, a reflecting plate arranged at the bottom of the lamp cover and a filter plate arranged outside the point light sources and fixed on the lamp cover, and the end face recognition LED lamp directly and vertically irradiates the end face of the retainer;

when the holder is judged, the brightness of the end face recognition LED lamp is not changed, an adaptive threshold value method is adopted, the threshold value of an image shot by a camera is set to be changed within an allowable range, the change range of the threshold value is defined as a standard threshold value +/-7, the standard threshold value is defined as a threshold value which is most suitable for determining a holder standard part during self-learning and is marked as T, the allowable adjustment range of the threshold value is 15 values in total, namely T-7, T-6, T-5, T-4, T-3, T-2, T-1, T, T +1, T +2, T +3, T +4, T +5, T +6 and T +7, when the holder image is judged, the end face of the holder is shot firstly, the threshold value is T, then the image is judged, the excircle diameter of the image is recognized, the recognized diameter is compared with the standard excircle diameter recorded by the standard part in the learning process, the following results occur:

a. if the deviation between the recognized excircle diameter and the standard excircle diameter is in a set range, the current threshold value is appropriate, and the subsequent judgment can be carried out;

b. if the recognized excircle diameter is smaller than the standard excircle diameter, the current threshold is set to be T-4, and then secondary photographing and recognition are carried out;

c. if the recognized excircle diameter is larger than the standard excircle diameter, the current threshold value is lower, the current threshold value is adjusted to be T +4, and then secondary photographing and recognition are carried out;

taking a picture for the second time and recognizing the picture according to the two conditions of b and c, comparing the recognized diameter of the second time with the standard excircle diameter recorded by the standard part in the learning process, obtaining the results a, b and c again, and repeating the steps until the deviation between the recognized excircle diameter and the standard excircle diameter is within a set range, so that the current threshold is proper; and if the deviation between the recognized outer circle diameter and the standard outer circle diameter is not in a set range all the time, the image is indicated to have a problem, the judgment is finished, and the retainer is indicated as a suspicious product to be manually processed.

The invention has the beneficial effects that:

1. the invention adopts a self-adaptive threshold value method, sets the threshold value of the image shot by the camera to be changed within an allowable range, and can realize the adjustment of 15 brightness levels by only adjusting the end face of the same retainer at most four times, thereby realizing the image judgment in the shortest time.

2. The method for shooting the self-adaptive threshold value of the picture by the camera is compared with the standard picture, so that the interference of image bright spots caused by the reflection of the surfaces of the workpiece and the tool on the feature recognition is effectively prevented, the requirement of the image recognition on a light source is effectively reduced, the influence of the light scattering on the image quality is reduced, and the misjudgment is effectively prevented.

Detailed Description

The invention will be further illustrated with reference to specific examples:

the invention relates to a self-adaptive method of a holder image threshold value suitable for distinguishing the front side and the back side of a holder, wherein the front side and the back side of the holder are distinguished by adopting an end face recognition LED lamp for illumination and adopting a camera for shooting, the camera is arranged right above the end face recognition LED lamp, and the axial lead of the camera and the axial lead of the end face recognition LED lamp are on the same straight line; the end face recognition LED lamp comprises an annular lamp cover, a ring of annular point light sources arranged inside the lamp cover, a reflecting plate arranged at the bottom of the lamp cover and a filter plate arranged outside the point light sources and fixed on the lamp cover, and the end face recognition LED lamp directly and vertically irradiates the end face of the retainer;

The above embodiments are only preferred embodiments of the present invention, and are not intended to limit the technical solutions of the present invention, so long as the technical solutions can be realized on the basis of the above embodiments without creative efforts, which should be considered to fall within the protection scope of the patent of the present invention.

Claims

1. A self-adaptive method of a retainer image threshold value suitable for judging the front and back of a retainer is characterized in that: the front side and the back side of the retainer are distinguished by adopting the end face recognition LED lamp to provide illumination, the camera is used for shooting, the camera is arranged right above the end face recognition LED lamp, and the axial lead of the camera and the axial lead of the end face recognition LED lamp are on the same straight line; the end face recognition LED lamp comprises an annular lamp cover, a ring of annular point light sources arranged inside the lamp cover, a reflecting plate arranged at the bottom of the lamp cover and a filter plate arranged outside the point light sources and fixed on the lamp cover, and the end face recognition LED lamp directly and vertically irradiates the end face of the retainer;