Crankshaft surface defect detection system based on machine vision

Abstract

The invention relates to a crankshaft surface defect detection system based on machine vision in the field of detection equipment, which is characterized in that a movement mechanism comprising a swinging assembly, a rotating assembly and a transverse moving assembly is arranged to transversely shoot and circumferentially shoot a crankshaft belonging to an axial complex polyhedron, so that a comprehensive defect detection image is obtained, and the defect omission ratio is reduced; meanwhile, when the defects are identified and positioned, rough positioning of the surface defects is obtained through transverse comprehensive shooting, then independent circumferential shooting is carried out on the recorded surface defects to obtain circumferential shooting images of the surface defects, and further identification and positioning are carried out on the surface defects; more accurate defect feature information and defect accurate positioning information are obtained, the accuracy of defect judgment is improved, and the accurate detection time is saved; in addition, through accurate defect positioning information, a marking mechanism is started to carry out marking pigment injection on the defects, and subsequent manual subsequent confirmation and corresponding treatment are facilitated.

Description

Technical Field

[0001] The invention relates to a surface defect detection system, in particular to a crankshaft surface defect detection system based on machine vision and applied in the field of detection equipment. Background Art

[0002] The crankshaft, usually composed of a crank and a journal connected to it, has a complex multi-faceted axial structure. The surface defect detection of the crankshaft has traditionally relied on manual operation. Given the complexity of its structure, the inspectors need to conduct detailed inspections from different angles. This process is highly dependent on the personal experience of the inspectors, which is not only inefficient, but also difficult to ensure the accuracy and consistency of the test results.

[0003] The existing patent with publication number CN113237895B discloses a metal surface defect detection system based on machine vision, which belongs to the field of metal machining and manufacturing technology. The invention includes a motion control system, an optical imaging system and an image processing system, and the motion control system, the optical imaging system and the image processing system are all connected to the industrial computer through their corresponding communication protocols. The motion control system includes a three-dimensional loading platform, and the three-dimensional loading platform adopts a gantry-type three-dimensional motion structure. The motion control system is used to place metal for metal surface defect detection. The present invention improves the optical detection means and optimizes the image processing algorithm, so that the detection accuracy is high, the detection speed is fast, the detection stability is high, the repeatability is good, and the output detection results are stable.

[0004] Although the above patent discloses a three-dimensional loading platform and improves the efficiency of photographing and inspecting complex components, when it is applied to the surface curve inspection of the crankshaft, an industrial camera is still required to photograph each surface of the crankshaft one by one and identify defects one by one. The movement of the industrial camera or the three-dimensional loading platform is complicated, and the inspection efficiency needs to be improved. Summary of the invention

[0005] In view of the above-mentioned prior art, the technical problem to be solved by the present invention is that the inspection efficiency of a crankshaft with a complex multi-faceted structure is low when it is inspected by a machine vision inspection system.

[0006] In order to solve the above problems, the present invention provides a crankshaft surface defect detection system based on machine vision, including a detection mechanism and a controller equipped with the detection system, the detection mechanism includes a base, a clamping frame for clamping the fixed crankshaft is fixedly connected to the base, an industrial camera is arranged on the outside of the clamping frame, the industrial camera is connected to a motion mechanism for driving it to move on the outside of the crankshaft and photograph various surfaces of the crankshaft, the motion mechanism includes a swinging component fixedly connected to the industrial camera and driving the industrial camera to swing laterally, the swinging component is provided with a first positioning sensor for detecting the swinging position of the industrial camera, the swinging component is fixedly connected to a rotating component for driving it to make a circular motion, the rotating component is provided with a second positioning sensor for detecting the rotational position of the industrial camera, the rotating component is connected to a transverse shift component for driving it to move laterally, and the transverse shift component is provided with a third positioning sensor for detecting the transverse shift position of the industrial camera; a marking mechanism fixedly connected to the swinging component is provided on one side of the industrial camera, the marking mechanism is used to spray marking pigment on the defective position;

[0007] The detection system includes an image acquisition module, the input end of the image acquisition module is connected to the industrial camera, the output end of the image acquisition module is connected to the defect recognition module, and the defect recognition module recognizes defects in the captured image; the output end of the defect recognition module is connected to the defect positioning module, the input end of the defect positioning module is also connected to the camera positioning module, the first positioning sensor, the second positioning sensor and the third positioning sensor are all connected to the input end of the camera positioning module, and the defect positioning module locates the position of the defect on the crankshaft according to the position of the defect in the captured image and the shooting position of the camera; the output end of the defect positioning module is connected to the control module, and the control module is also connected to the recording module, and the recording module records the defect position; the output end of the control module is also connected to the motion control module and the marking module, the swing component, the rotation component and the transverse movement component are all connected to the output end of the motion control module, and the marking mechanism is connected to the output end of the marking module.

[0008] In the above-mentioned crankshaft surface defect detection system based on machine vision, the detection efficiency of surface defects is improved and defects are marked by providing a motion mechanism and a detection system.

[0009] As a further improvement of the present application, the defect recognition module includes a feature recognition unit, a feature extraction unit, a category judgment unit and a feature storage unit; the feature recognition unit recognizes the image features in the horizontally captured images and the circularly captured images, and then the image feature extraction unit extracts the image features that meet the defect features to obtain the defect features of the surface defects, and finally the defect features of the surface defects are compared with the defect features of the irreparable defects in the feature storage unit through the category judgment unit to determine the defect type; the output end of the control module is also connected to an alarm module, and the output end of the alarm module is connected to a buzzer.

[0010] As a further improvement of the present application, the rotating assembly includes an outer ring and an inner ring rotatably connected to the inner wall of the outer ring, the swinging assembly is installed on the inner ring, the inner ring is fixedly connected to a gear ring, the gear ring is meshed with a driving gear, the driving gear is fixedly connected to the output shaft of the rotating motor, and the second position sensor is a second Hall sensor installed in the rotating motor.

[0011] As a further improvement of the present application, the outer ring and the inner ring are provided with matching notches, the longitudinal width of the notches is greater than the longitudinal width of the clamping frame, an arc-shaped groove is provided on the outer wall of the inner ring, the gear ring is fixedly connected to the inner wall of the arc-shaped groove, and the output shaft of the rotating motor passes through the outer ring and extends into the arc-shaped groove.

[0012] As a further improvement of the present application, the swing assembly includes a mounting plate for fixing the industrial camera, the mounting plate is fixedly connected to a swing rod on the side facing the inner ring, the swing rod extends into the inner ring and is rotatably connected thereto, an eccentric disk is clamped at the lower part of the swing rod, the eccentric disk is fixedly connected to the output shaft of the swing motor, and the first position sensor is a first Hall sensor arranged in the swing motor.

[0013] As a further improvement of the present application, the inner ring is provided with an equipment cavity for accommodating the swing motor, a swing groove connected to the equipment cavity is provided at the upper part, a hinged rod fixedly connected to the inner wall of the swing groove passes through the middle part of the swing rod, a rectangular through groove is provided at the lower part of the swing rod, the eccentric disk is a vertical disc-shaped structure and an eccentric column is fixedly connected at its circumferential edge, and the eccentric column is nested in the rectangular through groove.

[0014] As a further improvement of the present application, the transverse movement assembly includes a moving block fixedly connected to the upper end of the outer ring, the moving block is threadedly connected to a transverse screw rod, the transverse screw rod is rotatably connected to a support frame fixedly connected to the base, the end of the transverse screw rod is fixedly connected to the output shaft of the screw motor, and the third position sensor is a third Hall sensor arranged in the screw motor.

[0015] As a further improvement of the present application, the marking mechanism includes a pair of nozzles symmetrically arranged on both sides of the industrial camera and fixedly connected to the mounting plate, the lower ends of the pair of nozzles are respectively connected to injection pumps, the pair of injection pumps are respectively connected to storage bottles through pipelines, and the pair of storage bottles are respectively filled with marking pigments of different colors.

[0016] As a further improvement of the present application, the detection system includes the following steps when working:

[0017] Step 1, horizontal shooting: first, start the lateral movement component and the swing component through the motion control module, and start the industrial camera through the image acquisition module at the same time, so that the industrial camera moves from one end of the crankshaft to the other end along the axial parallel line direction of the crankshaft and swings and tilts when approaching the crank side wall during the lateral movement, so that the industrial camera performs the first horizontal shooting and obtains the first horizontal shooting image; second, when the first horizontal shooting is completed and the industrial camera moves to one end of the crankshaft, start the rotation component to rotate the industrial camera ninety degrees in the vertical plane, and then perform the second horizontal shooting to obtain the second horizontal shooting image; third, after the second horizontal shooting is completed, start the rotation component again, so that the industrial camera rotates ninety degrees in the same direction again, and then perform the third horizontal shooting to obtain the third horizontal shooting image; fourth, when the third horizontal shooting is completed, start the rotation component again to rotate the industrial camera ninety degrees in the same direction again, perform the fourth horizontal shooting, and obtain the fourth horizontal shooting image;

[0018] Step 2, roughly identify and locate the surface defects of the crankshaft: first, identify and extract the defect features in the first horizontal shooting image, the second horizontal shooting image, the third horizontal shooting image and the fourth horizontal shooting image through the feature recognition unit and the feature extraction unit in the defect recognition module, and then the category judgment unit compares the defect features of the surface defects with the defect features of the irreparable defects to obtain the similarity values ​​between the defect features of different surface defects and the defect features of the irreparable defects; second, the defect positioning module obtains the camera shooting position information and the image position information of the surface defects in the captured image according to the camera positioning module, and determines the defect spatial position information of the surface defects on the crankshaft; third, the defect positioning module transmits the obtained defect feature information, similarity value and defect spatial position information of the surface defects to the recording module; fourth, the recording module arranges and numbers the surface defects in order of their similarity values ​​from high to low;

[0019] Step 3, accurately identify and locate the surface defects of the crankshaft: first, the control module starts the motion mechanism through the motion control module based on the arrangement number of the surface defects in the recording module, so that the industrial camera moves to the outside of the surface defects one by one in the order of the arrangement number of the surface defects to take circular shots and obtain the circular shot images; second, the feature recognition unit, the feature extraction unit and the category judgment unit judge the type of the surface defects according to the circular shot images, and obtain the accurate positioning information of the surface defects through the defect positioning module;

[0020] Step four, different defects are marked with different pigments. When a surface defect is judged as an unrepairable defect by the category judgment unit, the rotating component drives the marking mechanism to move to a position relative to the surface defect according to the precise position information of the surface defect obtained in step three, and then the marking mechanism is started to spray the marking pigment, and the detection process is stopped. The buzzer light is activated through the alarm module to prompt the operator to replace the crankshaft; when the surface defect is judged as a repairable defect by the category judgment unit, the marking mechanism is started to spray another marking pigment, and then the circular shooting is continued in the order of the arrangement number of the surface defects.

[0021] In summary, the present invention, by providing a motion mechanism including a swinging component, a rotating component and a transverse displacement component, performs lateral and circumferential photography of a crankshaft which is an axially complex polyhedron, thereby obtaining a comprehensive defect detection image and reducing the defect missed detection rate; at the same time, when identifying and locating defects, a rough location of the surface defects is first obtained through lateral comprehensive photography, and then the recorded surface defects are photographed separately in a circular manner to obtain a circumferential photography image of the surface defects, and the surface defects are further identified and located to obtain more accurate defect feature information and defect precise positioning information, thereby improving the accuracy of defect judgment and saving precise detection time; in addition, through the defect precise positioning information, the marking mechanism is started to mark the defect with paint spray, which is convenient for subsequent manual confirmation and corresponding processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic diagram of the three-dimensional structure of the detection mechanism in this application;

[0023] Figure 2 It is a schematic diagram of the three-dimensional structure of the crank in this application;

[0024] Figure 3 It is a schematic diagram of the longitudinal cross-sectional structure of the detection mechanism in this application;

[0025] Figure 4 for Figure 3 A schematic diagram of the enlarged structure at A in the middle;

[0026] Figure 5 It is a schematic diagram of the exploded assembly of the rotating assembly in this application;

[0027] Figure 6 It is a schematic diagram of the three-dimensional structure of the marking mechanism in this application;

[0028] Figure 7 This is a schematic diagram of the modules of the detection system in the first embodiment of the present application;

[0029] Figure 8 This is a workflow diagram of the detection system in the first embodiment of the present application;

[0030] Fig. 9 This is a module schematic diagram of a defect recognition module in the second embodiment of the present application;

[0031] Fig.10 This is a schematic diagram of the modules of the detection system in the second embodiment of the present application;

[0032] Fig.11 This is a workflow diagram of the detection system in the second embodiment of the present application.

[0033] Description of the numbers in the figure:

[0034] 1. Base; 2. Clamping frame; 3. Industrial camera; 4. Swinging assembly; 401. Mounting plate; 402. Swinging rod; 403. Eccentric disk; 404. Swinging motor; 5. Rotating assembly; 501. Outer ring; 502. Inner ring; 503. Rotating motor; 6. Transverse movement assembly; 601. Moving block; 602. Transverse screw rod; 603. Support frame; 604. Screw rod motor; 7. Controller; 8. Marking mechanism; 801. Nozzle; 802. Jet pump; 803. Storage bottle; 9. Buzzer light. DETAILED DESCRIPTION

[0035] The following describes two implementation modes of the present application in detail with reference to the accompanying drawings.

[0036] The first implementation method:

[0037] Figure 1-8 The invention shows a crankshaft surface defect detection system based on machine vision, including a detection mechanism and a controller 7 equipped with the detection system. The detection mechanism includes a base 1, on which a clamping frame 2 for clamping a fixed crankshaft is fixedly connected, an industrial camera 3 is arranged on the outside of the clamping frame 2, the industrial camera 3 is connected with a motion mechanism for driving it to move outside the crankshaft and photograph various surfaces of the crankshaft, the motion mechanism includes a swinging assembly 4 fixedly connected to the industrial camera 3 and driving the industrial camera 3 to swing laterally, the swinging assembly 4 is provided with a first positioning sensor for detecting the swinging position of the industrial camera 3, the swinging assembly 4 is fixedly connected with a rotating assembly 5 for driving it to make a circular motion, the rotating assembly 5 is provided with a second positioning sensor for detecting the rotational position of the industrial camera 3, the rotating assembly 5 is connected with a transverse displacement assembly 6 for driving it to move laterally, and the transverse displacement assembly 6 is provided with a third positioning sensor for detecting the transverse displacement position of the industrial camera 3; a marking mechanism 8 fixedly connected to the swinging assembly 4 is arranged on one side of the industrial camera 3, and the marking mechanism 8 is used to spray marking pigment on the defective position;

[0038] See also Figure 7The detection system includes an image acquisition module, the input end of the image acquisition module is connected to the industrial camera 3, the output end of the image acquisition module is connected to the defect recognition module, and the defect recognition module recognizes defects in the captured image; the output end of the defect recognition module is connected to the defect positioning module, and the input end of the defect positioning module is also connected to the camera positioning module, the first positioning sensor, the second positioning sensor and the third positioning sensor are all connected to the input end of the camera positioning module, and the defect positioning module locates the position of the defect on the crankshaft according to the position of the defect in the captured image and the shooting position of the camera; the output end of the defect positioning module is connected to the control module, and the control module is also connected to the recording module, and the recording module records the defect position; the output end of the control module is also connected to the motion control module and the marking module, the swing component 4, the rotation component 5 and the lateral movement component 6 are all connected to the output end of the motion control module, and the marking mechanism 8 is connected to the output end of the marking module.

[0039] For details, please refer to Figure 1 , Figure 7 and Figure 8 , when the detection system works, it includes the following steps:

[0040] Step 1, horizontal shooting: first, start the lateral movement component 6 and the swing component 4 through the motion control module, and start the industrial camera 3 through the image acquisition module at the same time, so that the industrial camera 3 moves from one end of the crankshaft to the other end along the axial parallel line direction of the crankshaft and performs a swinging and tilting action when approaching the crank side wall during the lateral movement, so that the industrial camera 3 performs the first horizontal shooting and obtains the first horizontal shooting image; second, when the first horizontal shooting is completed and the industrial camera 3 moves to one end of the crankshaft, start the rotating component 5 to rotate the industrial camera 3 ninety degrees in the vertical plane, and then perform the second horizontal shooting to obtain the second horizontal shooting image; third, after the second horizontal shooting is completed, start the rotating component 5 again, so that the industrial camera 3 rotates ninety degrees in the same direction again, and then perform the third horizontal shooting to obtain the third horizontal shooting image; fourth, when the third horizontal shooting is completed, start the rotating component 5 again to rotate the industrial camera 3 ninety degrees in the same direction again, perform the fourth horizontal shooting, and obtain the fourth horizontal shooting image;

[0041] Specifically, the images of each surface of the entire crankshaft (the circumferential outer wall of both the journal and the crank, and the side walls of the crank) are quickly acquired through the lateral shooting and swing shooting of the industrial camera 3;

[0042] It should be noted that four horizontal shots are selected in the above step 1 to achieve coverage of all surfaces of the crankshaft. In actual use, the number of horizontal shots can be increased, but the rotation angle after each horizontal shot is not greater than 90 degrees, and the number of horizontal shots is equal to the rotation angle after each horizontal shot, which is equal to the integer quotient of 360 degrees and the number of horizontal shots. The more horizontal shots, the smaller the width of the horizontal shot image of the crankshaft, and the more details of the picture. However, too many shots are likely to cause the operation time to be too long and cumbersome. The operator selects the number of horizontal shots according to the needs;

[0043] Step 2, roughly identify and locate the surface defects of the crankshaft; first, identify the defects in the first horizontal shooting image, the second horizontal shooting image, the third horizontal shooting image and the fourth horizontal shooting image through the defect recognition module to obtain the defect feature information of the surface defects; second, the defect positioning module obtains the camera shooting position information when the surface defects are shot and the image position information of the surface defects in the horizontal shooting image according to the camera positioning module, and determines the defect spatial position information of the surface defects on the crankshaft. The defect positioning module transmits the obtained defect feature information and defect spatial position information of the surface defects to the recording module; third, the recording module arranges and numbers the surface defects from near to far according to the horizontal distance of the surface defects from the industrial camera 3 after the horizontal shooting is completed;

[0044] Specifically, by processing the horizontally shot images, the rough position of the surface defect can be quickly located, which is convenient for subsequent circumferential precise shooting;

[0045] It should be noted that defect characteristics include specification information such as length, width and shape of the defect;

[0046] Step 3, accurately identifying and locating the surface defects of the crankshaft: Based on the defect spatial position information and the arrangement number of the surface defects recorded by the recording module, the control module starts the lateral movement component 6 to move to the defect position through the motion control module and then starts the rotation component 5 and the swing component 4, so that the industrial camera 3 makes a circular motion and shoots one by one on the outside of the surface defects according to the arrangement number of the surface defects, and obtains a plurality of circular images of the surface defects. The defect recognition module and the defect positioning module accurately identify and locate the surface defects according to the circular images, and obtain accurate defect feature information and accurate positioning information of the surface defects;

[0047] Specifically, by circular shooting, the deformation of surface defects on each surface caused by angle problems during horizontal shooting can be overcome, and the precise defect characteristics and precise positions of surface defects can be obtained, which is convenient for subsequent judgment and marking;

[0048] Step 4, marking the surface defects with pigment: Based on the precise position information of the surface defects obtained in step 3, the rotating component 5 drives the marking mechanism 8 to move to a position relative to the surface defects, and then starts the marking mechanism 8 to spray the marking pigment to mark the position of the surface defects.

[0049] Specifically, the surface defects are accurately marked according to the precise positions of the surface defects obtained in step three.

[0050] Compared with the traditional surface defect detection system, the present invention is provided with a motion mechanism including a swing component 4, a rotation component 5 and a transverse displacement component 6, which can perform lateral and circumferential photography of the crankshaft which is an axially complex polyhedron, thereby obtaining a comprehensive defect detection image, reducing the defect missed detection rate, and improving the defect detection effect; at the same time, when identifying and locating defect features, the image position information of the surface defect and the camera position information when the camera photographs the defect are first obtained through the lateral shooting image, and the defect is roughly located, and then the recorded surface defect is photographed for the second time through circumferential photography to obtain a circumferential shooting image of the surface defect, and the surface defect is further identified and located, and more accurate defect feature information and defect precise positioning information are obtained, thereby improving the accuracy of defect judgment and saving precise detection time; in addition, through the defect precise positioning information, the marking mechanism 8 is started to mark the defect with paint spray, which is convenient for subsequent manual confirmation and corresponding processing.

[0051] See also Figure 4 and Figure 5 The rotating component 5 includes an outer ring 501 and an inner ring 502 rotatably connected to the inner wall thereof. The swinging component 4 is installed on the inner ring 502. The inner ring 502 is fixedly connected to a gear ring. The gear ring is meshed with a driving gear. The driving gear is fixedly connected to the output shaft of the rotating motor 503. The second position sensor is a second Hall sensor installed in the rotating motor 503.

[0052] Specifically, the rotating motor 503 drives the driving gear to rotate through the output shaft, the driving gear drives the gear ring meshing with it to rotate, the gear ring drives the inner ring 502 to rotate, and the inner ring 502 drives the swing assembly 4 and the industrial camera 3 installed on the swing assembly 4 to perform circular motion, thereby realizing the angle adjustment of the industrial camera 3 in the vertical plane; it should be noted that the second Hall sensor can monitor the rotation angle of the rotating motor 503. When the transmission ratio of the gear ring and the driving gear is known, the rotation angle of the industrial camera 3 in the vertical plane can be obtained, thereby determining the rotation position of the industrial camera 3 in the longitudinal vertical plane.

[0053] See also Figure 1 and Figure 5The outer ring 501 and the inner ring 502 are provided with notches that fit together, and the longitudinal width of the notches is greater than the longitudinal width of the clamping frame 2. An arc-shaped groove is provided on the outer wall of the inner ring 502, and the gear ring is fixedly connected to the inner wall of the arc-shaped groove. The output shaft of the rotating motor 503 passes through the outer ring 501 and extends into the arc-shaped groove.

[0054] Specifically, when the transverse moving assembly 6 drives the rotating assembly 5 to move transversely, the notches provided on the outer ring 501 and the inner ring 502 enable the rotating assembly 5 to move to the outside of the clamping frame 2, thereby facilitating the disassembly and assembly of the crankshaft.

[0055] See also Figure 4 The swing assembly 4 includes a mounting plate 401 for fixing the industrial camera 3, and the mounting plate 401 is fixedly connected with a swing rod 402 on the side facing the inner ring 502. The swing rod 402 extends into the inner ring 502 and is rotatably connected thereto. An eccentric disk 403 is clamped at the lower part of the swing rod 402, and the eccentric disk 403 is fixedly connected with the output shaft of the swing motor 404. The first position sensor is a first Hall sensor arranged in the swing motor 404.

[0056] Specifically, the output shaft of the swing motor 404 drives the eccentric disk 403 to rotate, the eccentric disk 403 drives the swing rod 402 to swing, and the swing rod 402 drives the mounting plate 401 and the industrial camera 3 mounted thereon to swing, thereby realizing the tilt of the industrial camera 3, and thereby realizing the shooting of the crank side wall on the crankshaft; it should be noted that the first Hall sensor can detect the rotation angle of the swing motor 404. Under the condition that the specifications of the eccentric disk 403 and the swing rod 402 are known, the swing position of the industrial camera 3 in the horizontal vertical plane can be obtained.

[0057] See also Figure 4 The inner ring 502 is provided with an equipment cavity for accommodating the swing motor 404, and a swing groove connected to the equipment cavity is provided at the upper part of the equipment cavity. A hinged rod fixedly connected to the inner wall of the swing groove passes through the middle of the swing rod 402, and a rectangular through groove is provided at the lower part of the swing rod 402. The eccentric disk 403 is a vertical disc-shaped structure and an eccentric column is fixedly connected at its circumferential edge, and the eccentric column is nested in the rectangular through groove.

[0058] Specifically, when the eccentric disc 403 drives the eccentric column to rotate, the eccentric column drives the swing rod 402 to swing, thereby achieving the adjustment of the tilt angle of the industrial camera 3 .

[0059] See also Figure 1The transverse movement assembly 6 includes a moving block 601 fixedly connected to the upper end of the outer ring 501, the moving block 601 is threadedly connected to a transverse screw rod 602, the transverse screw rod 602 is rotatably connected to a support frame 603 fixedly connected to the base 1, the end of the transverse screw rod 602 is fixedly connected to the output shaft of the screw motor 604, and the third position sensor is a third Hall sensor arranged in the screw motor 604.

[0060] Specifically, the output shaft of the screw motor 604 drives the transverse screw 602 to rotate, the transverse screw 602 drives the moving block 601 to move laterally, and the moving block 601 drives the rotating component 5 to move laterally. The third Hall sensor can detect the rotation angle of the screw motor 604. Under the condition that the specifications of the transverse screw 602 and the moving block 601 are known, the transverse position of the moving block 601 in the transverse plane can be obtained, and then the transverse position of the industrial camera 3 in the transverse plane can be obtained. The position information of the industrial camera 3 in space can be obtained by combining the position information of the first position sensor, the second position sensor and the third position sensor, and then the shooting position information of the industrial camera 3 when shooting defects can be obtained.

[0061] Second implementation method:

[0062] Figure 9-11 A crankshaft surface defect detection system based on machine vision is shown. On the basis of the first embodiment, the defect recognition module includes a feature recognition unit, a feature extraction unit, a category judgment unit and a feature storage unit; the feature recognition unit recognizes image features in the transversely shot image and the circularly shot image, and then the image feature extraction unit extracts image features that meet the defect features to obtain the defect features of the surface defects, and finally the category judgment unit compares the defect features of the surface defects with the defect features of the irreparable defects in the feature storage unit to determine the defect type; the output end of the control module is also connected to the alarm module, and the output end of the alarm module is connected to a buzzer 9.

[0063] It should be noted that the feature storage unit stores defect features of irreparable defects and defect features of repairable defects. The specific judgment criteria are set manually. For the convenience of explaining irreparable defects, an example is given. For example, if the shape of the defect feature of the extracted surface defect is close to an ellipse and the width is greater than 0.2 mm, then the surface defect is judged to be an irreparable defect.

[0064] For details, please refer to Figure 6 and Figure 9-11 , replace step 2, step 3 and step 4 in the first embodiment with the following steps 2, step 3 and step 4 respectively:

[0065] Step 2, roughly identify and locate the surface defects of the crankshaft: first, identify and extract the defect features in the first horizontal shooting image, the second horizontal shooting image, the third horizontal shooting image and the fourth horizontal shooting image through the feature recognition unit and the feature extraction unit in the defect recognition module, and then the category judgment unit compares the defect features of the surface defects with the defect features of the irreparable defects to obtain the similarity values ​​between the defect features of different surface defects and the defect features of the irreparable defects; second, the defect positioning module obtains the camera shooting position information and the image position information of the surface defects in the captured image according to the camera positioning module, and determines the defect spatial position information of the surface defects on the crankshaft; third, the defect positioning module transmits the obtained defect feature information, similarity value and defect spatial position information of the surface defects to the recording module; fourth, the recording module arranges and numbers the surface defects in order of their similarity values ​​from high to low;

[0066] Specifically, the rough position of the surface defect is obtained by horizontal shooting, and the surface defects are sorted according to the similarity with the irreparable defects, so as to improve the overall detection efficiency; in addition, it should be noted that in the first embodiment, the arrangement numbering of the surface defects is based on the distance of the surface defects from the industrial camera at the end of the horizontal shooting. In this embodiment, the arrangement numbering of the surface defects is based on the similarity between the surface defects and the irreparable defects, so as to further improve the overall detection efficiency (when the irreparable defects are found, it is not necessary to inspect other parts of the crankshaft);

[0067] Step 3, accurately identifying and locating the surface defects of the crankshaft: First, the control module starts the motion mechanism through the motion control module based on the arrangement number of the surface defects in the recording module, so that the industrial camera 3 moves to the outside of the surface defects one by one in the order of the arrangement number of the surface defects to perform circular shooting and obtain the circular shooting image; second, the feature recognition unit, the feature extraction unit and the category judgment unit judge the type of the surface defect according to the circular shooting image, and obtain the accurate positioning information of the surface defect through the defect positioning module;

[0068] Specifically, the defect characteristics and precise positions of surface defects are further obtained through circumferential photography to improve the accuracy of judgment;

[0069] It should be noted that when judging the type of surface defects, since the circumferential photographed image of the surface defects is obtained, the feature recognition unit and the feature extraction unit further obtain more accurate defect feature information of the surface defects, and compare the accurate defect feature information of the surface defects with the defect feature information of the irreparable defects in the feature storage unit, and then accurately judge whether the current surface defect is an irreparable defect;

[0070] Step four, different defects are marked with different pigments. When a surface defect is judged as an unrepairable defect by the category judgment unit, the rotating component 5 drives the marking mechanism 8 to move to a position relative to the surface defect according to the precise position information of the surface defect obtained in step three, and then the marking mechanism 8 is started to spray the marking pigment, and the detection process is stopped. The buzzer light 9 is started through the alarm module to prompt the operator to replace the crankshaft; when the surface defect is judged as a repairable defect by the category judgment unit, the marking mechanism 8 is started to spray another marking pigment, and then the circular shooting is continued in the order of the arrangement number of the surface defects.

[0071] Specifically, the defect characteristics of surface defects are accurately acquired through circular photography, and the defect category is accurately determined, thereby saving overall inspection time (when an unrepairable defect is found, there is no need to perform circular photography inspection on other parts of the crankshaft).

[0072] Compared with the traditional surface defect detection system, the present invention is provided with a defect recognition module including a feature recognition unit, a feature extraction unit, a category judgment unit and a feature storage unit. When shooting horizontally, the surface defects with high similarity in defect features to the irreparable defects are sorted, and when shooting circularly, the surface defects are photographed one by one according to the similarity value from high to low, and the surface defects suspected of being irreparable defects are detected first, thereby improving the overall detection efficiency, marking the irreparable defects and issuing a shutdown alarm, which is convenient for manual confirmation and processing, further improving the detection efficiency, and saving the time of batch inspection of crankshafts.

[0073] See also Figure 6 The marking mechanism 8 includes a pair of nozzles 801 symmetrically arranged on both sides of the industrial camera 3 and fixedly connected to the mounting plate 401. The lower ends of the pair of nozzles 801 are respectively connected to injection pumps 802, and the pair of injection pumps 802 are respectively connected to storage bottles 803 through pipelines. The pair of storage bottles 803 are respectively filled with marking pigments of different colors.

[0074] Specifically, when marking surface defects, repairable defects and non-repairable defects are marked with marking pigments of different colors for easy identification and distinction.

[0075] In view of current practical needs, the above-mentioned implementation mode adopted in this application is not limited to the scope of protection. Various changes made within the knowledge scope of technical personnel in this field without departing from the concept of this application still fall within the scope of protection of the present invention.

Claims

1. The crankshaft surface defect detection system based on machine vision is characterized by: The invention comprises a detection mechanism and a controller (7) equipped with a detection system, wherein the detection mechanism comprises a base (1), a clamping frame (2) for clamping a fixed crankshaft is fixedly connected to the base (1), an industrial camera (3) is arranged outside the clamping frame (2), the industrial camera (3) is connected to a motion mechanism for driving the industrial camera (3) to move outside the crankshaft and photograph each surface of the crankshaft, the motion mechanism comprises a swinging assembly (4) fixedly connected to the industrial camera (3) and driving the industrial camera (3) to swing horizontally, the swinging assembly (4) is provided with a first A positioning sensor, wherein the swing component (4) is fixedly connected to a rotating component (5) that drives it to perform circular motion, the rotating component (5) is provided with a second positioning sensor for detecting the rotational position of the industrial camera (3), the rotating component (5) is connected to a transverse movement component (6) that drives it to move transversely, the transverse movement component (6) is provided with a third positioning sensor for detecting the transverse movement position of the industrial camera (3); a marking mechanism (8) fixedly connected to the swing component (4) is provided on one side of the industrial camera (3), the marking mechanism (8) is used to spray marking pigment on the defective position; The detection system comprises an image acquisition module, the input end of the image acquisition module is connected to the industrial camera (3), the output end of the image acquisition module is connected to the defect recognition module, and the defect recognition module recognizes defects in the captured image; the output end of the defect recognition module is connected to the defect positioning module, the input end of the defect positioning module is also connected to the camera positioning module, the first positioning sensor, the second positioning sensor and the third positioning sensor are all connected to the input end of the camera positioning module, and the defect positioning module locates the position of the defect on the crankshaft according to the position of the defect in the captured image and the shooting position of the camera; the output end of the defect positioning module is connected to the control module, and the control module is also connected to the recording module, and the recording module records the defect position; the output end of the control module is also connected to the motion control module and the marking module, the swing component (4), the rotation component (5) and the lateral movement component (6) are all connected to the output end of the motion control module, and the marking mechanism (8) is connected to the output end of the marking module.

2. The crankshaft surface defect detection system based on machine vision according to claim 1 is characterized in that: The defect recognition module comprises a feature recognition unit, a feature extraction unit, a category judgment unit and a feature storage unit; the feature recognition unit recognizes image features in the transversely shot image and the circumferentially shot image, and then the image feature extraction unit extracts image features that meet the defect features to obtain defect features of the surface defect, and finally the category judgment unit compares the defect features of the surface defect with the defect features of the irreparable defect in the feature storage unit to determine the defect type; the output end of the control module is also connected to an alarm module, and the output end of the alarm module is connected to a buzzer (9).

3. The crankshaft surface defect detection system based on machine vision according to claim 1, characterized in that: The rotating assembly (5) comprises an outer ring (501) and an inner ring (502) rotatably connected to the inner wall of the outer ring (501), the swinging assembly (4) is mounted on the inner ring (502), the inner ring (502) is fixedly connected to a gear ring, the gear ring is meshed with a driving gear, the driving gear is fixedly connected to the output shaft of the rotating motor (503), and the second position sensor is a second Hall sensor mounted in the rotating motor (503).

4. The machine vision-based crankshaft surface defect detection system according to claim 3, characterized in that: The outer ring (501) and the inner ring (502) are provided with notches that fit together, the longitudinal width of the notches being greater than the longitudinal width of the clamping frame (2), an arc-shaped groove being provided on the outer wall of the inner ring (502), the gear ring being fixedly connected to the inner wall of the arc-shaped groove, and the output shaft of the rotating motor (503) passing through the outer ring (501) and extending into the arc-shaped groove.

5. The crankshaft surface defect detection system based on machine vision according to claim 1, characterized in that: The swing assembly (4) comprises a mounting plate (401) for fixing the industrial camera (3); a swing rod (402) is fixedly connected to the mounting plate (401) on the side facing the inner ring (502); the swing rod (402) extends into the inner ring (502) and is rotatably connected thereto; an eccentric disk (403) is clamped at the lower part of the swing rod (402); the eccentric disk (403) is fixedly connected to the output shaft of the swing motor (404); and the first position sensor is a first Hall sensor arranged in the swing motor (404).

6. The machine vision-based crankshaft surface defect detection system according to claim 5, characterized in that: The inner ring (502) is provided with an equipment cavity for accommodating the swing motor (404), the upper part of the equipment cavity is provided with a swing groove connected thereto, the middle part of the swing rod (402) is penetrated by a hinge rod fixedly connected to the inner wall of the swing groove, the lower part of the swing rod (402) is provided with a rectangular through groove, the eccentric disk (403) is a vertical disc-shaped structure and an eccentric column is fixedly connected to its circumferential edge, and the eccentric column is nested in the rectangular through groove.

7. The crankshaft surface defect detection system based on machine vision according to claim 1, characterized in that: The transverse movement assembly (6) comprises a moving block (601) fixedly connected to the upper end of the outer ring (501); the moving block (601) is threadedly connected to a transverse screw rod (602); the transverse screw rod (602) is rotatably connected to a support frame (603) fixedly connected to the base (1); the end of the transverse screw rod (602) is fixedly connected to the output shaft of the screw motor (604); and the third position sensor is a third Hall sensor arranged in the screw motor (604).

8. The machine vision-based crankshaft surface defect detection system according to claim 2, characterized in that: The marking mechanism (8) comprises a pair of nozzles (801) symmetrically arranged on both sides of the industrial camera (3) and fixedly connected to the mounting plate (401); the lower ends of the pair of nozzles (801) are respectively connected to injection pumps (802); the pair of injection pumps (802) are respectively connected to storage bottles (803) through pipelines; and the pair of storage bottles (803) are respectively filled with marking pigments of different colors.

9. The machine vision-based crankshaft surface defect detection system according to claim 8, characterized in that: The detection system includes the following steps when working: Step 1: Shoot horizontally: Step 1.1, start the lateral movement component (6) and the swing component (4) through the motion control module, and start the industrial camera (3) through the image acquisition module, so that the industrial camera (3) moves from one end of the crankshaft to the other end along the axial parallel line direction of the crankshaft and performs a swinging and tilting action when approaching the crank side wall during the lateral movement, so that the industrial camera (3) performs a lateral shooting and obtains a lateral shooting image; Step 1.2, when a horizontal shooting is completed and the industrial camera (3) moves to one end of the crankshaft, the rotating assembly (5) is started to rotate the industrial camera (3) ninety degrees in the vertical plane, and then the horizontal shooting is continued to obtain another horizontal shooting image; Step 1.3, repeat step 1.2 twice, and obtain four horizontal images in different directions in total; Step 2, roughly identifying and locating the surface defects of the crankshaft: First, the defect features in the four horizontally shot images are identified and extracted by the feature recognition unit and the feature extraction unit in the defect recognition module, and then the category judgment unit compares the defect features of the surface defects with the defect features of the irreparable defects to obtain the similarity values ​​of the defect features of different surface defects and the defect features of the irreparable defects; Second, the defect location module obtains the camera shooting position information and the image position information of the surface defect in the captured image according to the camera positioning module, and determines the defect spatial position information of the surface defect on the crankshaft; Third, the defect location module transmits the obtained defect feature information, similarity value and defect spatial position information of the surface defect to the recording module. Fourth, the recording module arranges and numbers the surface defects in descending order according to their similarity values. Step 3, accurately identifying and locating the surface defects of the crankshaft: first, the control module starts the motion mechanism through the motion control module based on the arrangement number of the surface defects in the recording module, so that the industrial camera (3) moves to the outside of the surface defects one by one in the order of the arrangement number of the surface defects to perform circular shooting, and obtains the circular shooting image; second, the feature recognition unit, the feature extraction unit and the category judgment unit judge the type of the surface defect according to the circular shooting image, and obtain the accurate positioning information of the surface defect through the defect positioning module; Step 4: different defects are marked with different pigments. When a surface defect is judged as an unrepairable defect by the category judgment unit, the rotating component (5) drives the marking mechanism (8) to move to a position relative to the surface defect according to the precise position information of the surface defect obtained in step 3, and then the marking mechanism (8) is started to spray the marking pigment, and the detection process is stopped. The buzzer light (9) is started through the alarm module to prompt the operator to replace the crankshaft; When the surface defect is judged as a repairable defect by the category judgment unit, the marking mechanism (8) is started to spray another marking pigment, and then the circumferential shooting is continued in the order of the arrangement number of the surface defects.

Images