CN104101608A - Intelligent detecting device for detecting defects of multi-type irregularly shaped product - Google Patents

Abstract

The invention discloses an intelligent detection device capable of detecting defects of multi-type irregular-shaped products, including a glass slide table, a bottleneck feeding mechanism and an image acquisition system. The glass slide table is arranged on a frame inclined relative to the horizontal plane; The feeding mechanism includes two scoop-shaped baffles with the same structural size, and the two scoop-shaped baffles are arranged symmetrically; the image acquisition system includes a top-view image acquisition mechanism, a side-view image acquisition mechanism, an image acquisition trigger mechanism and an image analysis system. . The intelligent detection device provided by the present invention, which can detect defects of multi-type irregular-shaped products, can detect various types of products without a preset database; at the same time, the design of the camera and lights in the present invention also enables the shooting of products The effect is better, and the detection quality is improved; the present invention basically realizes the automation and intelligence of multi-type irregular product defect detection, greatly saves manpower and material resources, and improves detection accuracy.

Description

An intelligent detection device that can perform defect detection on multi-model irregular-shaped products

technical field

The invention relates to an intelligent detection device capable of detecting defects of multi-model irregular-shaped products, which belongs to the product quality detection technology.

Background technique

Defect detection of production line products is one of the important links before factory products leave the factory. Manual inspection is difficult to ensure sufficient detection accuracy due to defects such as slow detection speed, human fatigue and emotional factors. Therefore, introducing machine vision technology into the product inspection process, using machines instead of human eyes to identify, judge and measure target objects, and replacing traditional manual inspection methods will help improve inspection efficiency and industrial automation.

Machine vision technology is a kind of visual science and engineering. Since the development of this technology in the 1950s, it has made great progress in the depth and breadth of research, and its application in industrial automation has become more and more extensive. Among them There is also no shortage of examples of this technique being applied to product inspection. However, most of the traditional machine vision product inspection systems have restrictions on the inspection products, that is, a set of inspection systems can only inspect one product and one model, or can only inspect several product types and models already in the inspection system database. detection. The former means that a diversified manufacturer needs to be equipped with multiple sets of testing systems and production lines to complete the testing of all products, which will cause a certain amount of production waste and increase the production cost of the factory. The latter detection system means that every time a product model is added or replaced, the database of the system needs to be edited again, and the system test is carried out here. Obviously, this cannot adapt to the faster and faster pace of product replacement in today's society. With the updating and upgrading of products, it will cost manpower and material resources to update the detection system. At the same time, changes to its camera parameters and lighting are often still done manually. All in all, the above two traditional detection systems do not have the characteristics of intelligence.

Contents of the invention

Purpose of the invention: Aiming at the technical and application defects of the traditional production line product detection system, the present invention provides an intelligent detection device capable of detecting defects of multi-type irregular-shaped products, which can automatically identify standard parts and store the characteristics of standard parts Information, can automatically adjust the assembly line structure and intelligently detect products.

Technical scheme: in order to achieve the above object, the technical scheme adopted in the present invention is:

An intelligent detection device that can perform defect detection on multi-model irregular-shaped products, including a glass slide table, a bottleneck feeding mechanism and an image acquisition system;

The glass slide is installed on the frame inclined relative to the horizontal plane, a vertical plane S is defined in the slope direction of the glass slide, and the vertical plane S divides the glass slide into left and right regions;

The bottleneck feeding mechanism includes two scoop-shaped baffles with the same structural size. The scoops of the two scoop-shaped baffles are set opposite to and symmetrical to the vertical plane S. The scoop-shaped baffles are arranged on the glass slide table to limit the The movement area of the test piece on the glass slide table; the position of the spoon mouth of the scoop-shaped baffle is higher than the position of the spoon handle, and the ends of the two spoon mouths are used as the feeding end of the piece to be tested. The distance between the deep recesses of the two spoon mouths is the smallest and ensures Detect the passage of the product, and the tails of the two spoon handles converge;

The image acquisition system includes a top view image acquisition mechanism, a side view image acquisition mechanism, an image acquisition trigger mechanism and an image analysis system, and the image acquisition areas of the top view image acquisition mechanism and the side view image acquisition mechanism are located in the area where the two spoon handles are opposite; The bird's-eye view image acquisition mechanism includes a bird's-eye view camera, which is arranged vertically on a glass slide and is located on a vertical plane S, where the head-down camera is located above the glass slide, and a plane light source is arranged directly below the glass slide to face the head-down camera; The side view image acquisition mechanism includes a first camera, a second camera and a third camera, the first camera, the second camera and the third camera are placed in a plane parallel to the glass slide, and the top view camera is placed on the glass slide The projection is the center, 120 ° interval ring setting, corresponding to the first camera is provided with a first point light source, corresponding to the second camera is provided with a second point light source, corresponding to the third camera is provided with a third point light source; the image acquisition trigger mechanism It includes an optical fiber sensor, the optical fiber sensor is arranged directly against the vertical plane S, and simultaneously triggers the overhead camera, the first camera, the second camera and the third camera to work through the optical fiber sensor, and sends the collected image to the image analysis system; the image analysis The system includes a standard part image acquisition unit and a to-be-inspected part image acquisition unit. The image collected by the to-be-inspected part image acquisition unit is compared with the image collected by the standard part image acquisition unit to determine whether the to-be-inspected part is a qualified product.

The present invention changes the width of the assembly line of the traditional detection system according to the requirements of product size limitation, and adopts two scoop-shaped baffles to form the slideway of the product. The spacing can be designed according to the width of the assembly line; the narrow area between the deep recesses of the two scoops is used to adjust the position where the product enters the photo area, and the spacing can be designed according to the shape and size of the product. When the product enters the slide at a certain initial speed and under the action of gravity, it enters the camera area with a certain posture after being adjusted through the narrow area between the deep recesses of the two scoop mouths. Before entering the camera area, the product will first be detected by the optical fiber sensor It is sensed that the fiber optic sensor triggers four cameras to work according to the sensing information; since there is a certain delay in triggering the camera by the fiber optic sensor, considering the delay and the continuous movement of the product, the distance between the deep recesses of the two spoon mouths and the position of the fiber optic sensor can be comprehensively designed. When the camera is working, the product is located in the center of the side-view image acquisition mechanism as much as possible; at the same time, the focal length and position of the four cameras are adjusted to make the images captured by the cameras the clearest and most recognizable.

In specific work, after adjusting each parameter, first take a picture of the standard part, and the standard part images captured by the four cameras are sent to the standard part image acquisition unit as a reference image; in order to improve the accuracy of judgment, it is necessary to More standard part images, generally speaking, at least five standard part images are required. After the image of the standard part is collected, the part to be tested can be detected. Based on the same parameter conditions as the standard part, the photo of the part to be tested is taken, and the obtained image of the part to be tested is sent to the image acquisition unit of the part to be tested. The image of the inspected part is compared with the image of the standard part to analyze whether the part to be inspected is qualified or not.

In this case, there may not be any database in the image analysis system. Before testing a certain batch of products, several standard parts pass through the slideway in turn, and the images of the standard parts are obtained by the camera and then transmitted to the image analysis system. Then the image analysis system based on The images of these standard parts obtained are used for the detection of this batch of products; this means that any product that meets the detection requirements can use this device, which has a large detection range; therefore, the device and corresponding methods given in this case can be used without There are restrictions on the types of products, and there is even no need to expand and modify the database. Correspondingly, according to the acquired image information of the standard part, the robot arm can be actively or controlled to initially adjust the relevant parameters, and when the adjustment is completed, the relevant image acquisition starts.

Preferably, the inclination angle of the glass slide table relative to the horizontal plane is 30°.

Preferably, displacement sensors are installed at the deep recesses of the two spoon mouths, and the distance between the deep recesses of the two spoon mouths is detected by the displacement sensors, that is, the width of the bottleneck position is measured.

Preferably, both the head and the tail of the two scoop-shaped baffles are installed on the frame through a screw slide mechanism, and the distance between the two scoop-shaped baffles is adjusted by adjusting the screw slide mechanism.

Preferably, the overlooking camera is installed on the frame through the lifting camera frame, and the lifting camera frame includes a ball screw mechanism, a crossbeam and a stepping motor, and the leading screw of the ball screw mechanism is driven by the stepping motor to move vertically up and down, and the crossbeam One end of the beam is fixed to the lead screw of the ball screw mechanism, and the other end of the beam is installed with a top-view camera.

Preferably, the top-view camera, the first camera, the second camera and the third camera are all provided with automatic focusing devices.

Beneficial effects: the intelligent detection device provided by the present invention, which can detect defects of multi-type irregular-shaped products, can detect various types of products without a preset database; at the same time, the design of the camera and lights in the present invention also enables The shooting effect of the product is better, and the detection quality is improved; the present invention basically realizes the automation and intelligence of multi-type irregular product defect detection, greatly saves manpower and material resources, and improves the detection accuracy.

Description of drawings

Fig. 1 is the top view structure schematic diagram of the present invention;

Fig. 2 is the three-dimensional structure schematic diagram of the present invention;

Fig. 3 is a working flow diagram of the present invention.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings.

As shown in Fig. 1 and Fig. 2, it is an intelligent detection device capable of detecting defects of multi-type irregular-shaped products, including a glass slide table 1, a bottleneck feeding mechanism, an image acquisition system and an industrial computer.

The glass slide 1 is installed on the frame with an inclination angle of 30° relative to the horizontal plane. A vertical plane S is defined in the direction of the slope of the glass slide 1, and the vertical plane S divides the glass slide 1 into two left and right sides. area.

The bottleneck feeding mechanism includes two scoop-shaped baffles 2 with the same structural size. The spoon openings of the two scoop-shaped baffles 2 are set opposite to and symmetrical to the vertical plane S. The scoop-shaped baffles 2 are arranged on the glass slide 1 above, to limit the movement area of the piece to be tested on the glass slide 1; The distance between two spoon-shaped baffles 2; the position of the mouth of the spoon-shaped baffle 2 is higher than the position of the handle of the spoon, and the ends of the two spoons are used as the feeding end of the piece to be tested. The distance between the deep recesses of the two spoons is the smallest and ensures that only One detection product passes, and the tails of the handles of the two spoons converge; the deep recesses of the two spoon mouths are provided with displacement sensors 16, and the distance between the deep recesses of the two spoon mouths is detected by the displacement sensor 16.

The image collection system includes a top view image collection mechanism, a side view image collection mechanism, an image collection trigger mechanism and an image analysis system, and the image collection areas of the top view image collection mechanism and the side view image collection mechanism are located in the area where the handles of the two spoons face each other.

The bird's-eye view image collection mechanism includes a bird's-eye view camera 4, which is arranged vertically to the glass slide 1 and is located on a vertical plane S, and the bird's-eye view camera 4 is located above the glass slide 1, facing the bird's-eye view camera below the glass slide 1 4. A plane light source 9 is provided; the side-view image acquisition mechanism includes a first camera 5 installed on a first camera bracket 8, a second camera 6 installed on a second camera bracket 13, and a third camera bracket 14 The third camera 7 on the top, the first camera 5, the second camera 6 and the third camera 7 are placed in a plane parallel to the glass slide 1, and take the projection of the overhead camera 4 on the glass slide 1 as the center, 120 ° interval ring arrangement, corresponding to the first camera 5 is provided with a first point light source 10, corresponding to the second camera 6 is provided with a second point source of light 11, corresponding to the third camera 7 is provided with a third point source of light 12; The first camera 5 , the second camera 6 and the third camera 7 are all provided with an automatic focusing device 18 .

Described bird's-eye view camera 4 is installed on the frame by elevating camera frame 3, and elevating camera frame 3 comprises ball screw mechanism, crossbeam and stepper motor, drives the leading screw of ball screw mechanism to move vertically by stepper motor, crossbeam One end of the crossbeam is fixed with the leading screw of the ball screw mechanism, and the other end of the crossbeam is installed with a bird's-eye view camera 4.

Described image acquisition triggering mechanism comprises fiber optic sensor 15, and fiber optic sensor 15 is set up against vertical plane S, triggers overlooking camera 4, first camera 5, second camera 6 and the 3rd camera 7 work simultaneously by fiber optic sensor 15, and will The collected image is sent to the image analysis system; the image analysis system includes a standard part image acquisition unit and a to-be-inspected part image acquisition unit, and the image collected by the to-be-tested part image acquisition unit is compared with the image collected by the standard part image acquisition unit, Determine whether the part to be tested is a qualified product.

The description of each part is as follows:

Glass slide 1: The friction force is small, and products with an initial velocity can slide down naturally only under the action of gravity, and the transparent material of glass slide 1 does not hinder the illumination of the plane light source 9 .

Lifting camera frame 3: According to different product models, the stepping motor can drive the ball screw to move, adjust the height of the beam, and realize the purpose of intelligently adjusting the height of the overlooking camera 4, so that the size of the product in the camera field of view is appropriate; at the same time, the side The video image acquisition mechanism can also adopt a similar structure to achieve the purpose of intelligent adjustment of the distance between the camera and the center point.

Displacement sensor 16: A laser displacement sensor is used to measure the width of the bottleneck and feed it back to the industrial computer. The industrial computer intelligently adjusts the width of the bottleneck to an appropriate distance according to the captured image information.

Planar light source 9: LED flat light is used, which can be produced in any size, has sufficient brightness and light stability, and can cover the photographing area of the corresponding camera.

Optical fiber sensor 15: its setting position can be adjusted according to relevant information such as product size, and is used to trigger the camera.

Lead screw slide mechanism 17: adjust the distance between the scoop-shaped baffles 2 on both sides according to the information given by the industrial computer.

Automatic focusing device 18: used to adjust the definition of the camera, and determine the rotation angle of the stepping motor according to the signal given by the industrial computer, thereby driving the fine-tuning gear coaxially fixed with the camera lens to achieve the purpose of automatic focusing.

Industrial computer: placed under the glass slide table 1, with a built-in image analysis system, which receives the detection signals of all testing equipment, and after the industrial computer makes judgments based on the obtained image information of standard parts, the controller automatically drives a series of motors or mechanical fine-tuning hands Initial adjustments such as camera focal length, height, optical fiber sensor position, and arc controller adjustment are performed without manual adjustment; when the adjustment is completed, the batch of parts is inspected, and the inspection information is analyzed to obtain the inspection results.

In this case, the planar light source 9 is back-facing planar light, and is installed under the glass slide 1; the overlooking camera 4 is located above the glass slide 1, and the lens is facing the glass slide 1, so the backlighting method is adopted; the first The point light source 10, the second point light source 11 and the third point light source 12 are arranged on the sides of the lens of the first camera 5, the second camera 6 and the third camera 7, and the method of frontal lighting with ring light is adopted for taking pictures on the side. For the detection of product shape defects, compared with the possibility of shadows and image judgment errors caused by the front lighting method, the backlighting method is used in this case. This lighting method will cause the tested product to form dark shadows on the bright background. , has a very obvious contrast, so it is very suitable for the detection of the target outline; the side and the front are illuminated to obtain information such as the height, hollowing out, and color of the product; this device can realize the shape defect detection and size detection of various irregular products.

The unqualified products detected by the device in this case need to be rejected by the rejecting device to achieve further intelligence; a rejecting device is provided in this case, which is installed at the end of the glass slide 1, and the main device is an electromagnetic lock; The principle of electromagnetism, when the current is generated, the electromagnetic lock will generate a strong suction to straighten the tape, causing the baffle to generate an angle and guide the parts to go; on the contrary, after the power is turned off, the electromagnetic lock loses suction and the tape is retracted, and then As a result, the baffle is angled again, leading to another direction of the part.

The working process of the device in this case is shown in Figure 3, specifically: first, run the camera software, place the standard parts of this batch at the center of the camera’s field of view, and determine the imaging resolution and the position and proportion of the field of view occupied by the control panel. The controller automatically adjusts the focal length of the four cameras, the height of the elevating camera frame 3, and the positions of the scoop-shaped baffle 2 and the optical fiber sensor 15; runs the operating software, places about 5 standard parts under the camera, collects standard part information, standard part information After the collection is completed, the inspection of this batch of parts can be run; the parts are sent to the top of the glass slide 1 by the conveyor belt, and under the action of gravity, the parts pass through the bottleneck position of the spoon-shaped baffle 2 and then enter the center of the camera's field of view; the parts continue to slide down , the camera is triggered by the optical fiber sensor 15, and the four cameras take pictures at the same time to obtain a front view of the part and three side views of the part; the image analysis system analyzes the four pictures to obtain the detection result; the controller controls the left and right sides of the rejecting device according to the detection result Move to complete the diversion.

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 are also possible. It should be regarded as the protection scope of the present invention.

Claims (6)

1. An intelligent detection device capable of detecting defects of multi-type irregular-shaped products, characterized in that it includes a glass slide table (1), a bottleneck feeding mechanism and an image acquisition system; The glass slide (1) is installed on the frame obliquely relative to the horizontal plane, and a vertical plane S is defined in the slope direction of the glass slide (1), and the vertical plane S divides the glass slide (1) into left and right halves. regions; The bottleneck feeding mechanism includes two scoop-shaped baffles (2) with the same structural size. It is installed on the glass slide (1) to limit the movement area of the object to be tested on the glass slide (1); the position of the scoop mouth of the scoop-shaped baffle (2) is higher than the position of the spoon handle, and the ends of the two scoops are used as At the feeding end of the piece to be tested, the distance between the deep recesses of the two spoon mouths is the smallest to ensure that only one test product passes through, and the tails of the two spoon handles converge; The image acquisition system includes a top view image acquisition mechanism, a side view image acquisition mechanism, an image acquisition trigger mechanism and an image analysis system, and the image acquisition areas of the top view image acquisition mechanism and the side view image acquisition mechanism are located in the area where the two spoon handles are opposite; The bird's-eye view image acquisition mechanism includes a bird's-eye view camera (4), the head-view camera (4) is arranged vertically on the glass slide (1) and is positioned on the vertical plane S, and the bird's-eye view camera (4) is located above the glass slide (1), on the glass slide (1). The bottom of the slide table (1) is facing the top-view camera (4) and is provided with a plane light source (9); the side-view image acquisition mechanism includes a first camera (5), a second camera (6) and a third camera (7) , the first camera (5), the second camera (6) and the third camera (7) are placed in a plane parallel to the glass slide (1), and the top-view camera (4) is on the glass slide (1) The projection on the center is arranged in a ring at intervals of 120°, a first point light source (10) is set corresponding to the first camera (5), a second point light source (11) is set corresponding to the second camera (6), and a second point light source (11) is set corresponding to the third camera (7) be provided with the 3rd point light source (12); Described image acquisition triggering mechanism comprises optical fiber sensor (15), and optical fiber sensor (15) is set up against vertical plane S, triggers overlooking camera (15) simultaneously by optical fiber sensor (15) 4), the first camera (5), the second camera (6) and the third camera (7) work, and the images collected are sent to the image analysis system; the image analysis system includes a standard part image acquisition unit and a A part image acquisition unit is used to compare the image collected by the image acquisition unit of the part to be inspected with the image collected by the image acquisition unit of the standard part to judge whether the part to be inspected is a qualified product. 2. The intelligent detection device capable of detecting defects of multi-type irregular-shaped products according to claim 1, characterized in that: the inclination angle of the glass slide table (1) relative to the horizontal plane is 30°. 3. The intelligent detection device according to claim 1, which can detect defects of multi-type irregular-shaped products, characterized in that: displacement sensors (16) are arranged at the deep recesses of the two scoop mouths, and the displacement sensors (16) 16) Detect the distance between the deep recesses of the two spoon mouths. 4. The intelligent detection device capable of detecting defects of multi-type irregular-shaped products according to claim 1, characterized in that: the head and tail of the two spoon-shaped baffles (2) pass through the screw slide mechanism (17) be installed on the frame, adjust the distance between two scoop-shaped baffles (2) by adjusting the lead screw slide mechanism (17). 5. The intelligent detection device capable of detecting defects of multi-type irregular-shaped products according to claim 1, characterized in that: the overhead camera (4) is installed on the frame by lifting the camera frame (3), and the lifting Camera frame (3) comprises ball screw mechanism, crossbeam and stepper motor, the leading screw of ball screw mechanism is driven by stepper motor to move vertically up and down, and one end of crossbeam is fixed with the leading screw of ball screw mechanism, and the other end of crossbeam A top view camera (4) is installed at one end. 6. The intelligent detection device capable of detecting defects of multi-type irregular-shaped products according to claim 1, characterized in that: the overlooking camera (4), the first camera (5), and the second camera (6) An automatic focusing device (18) is arranged on the third camera (7).