CN112304964A - Fuse box vision detection system - Google Patents
Fuse box vision detection system Download PDFInfo
- Publication number
- CN112304964A CN112304964A CN202011378565.5A CN202011378565A CN112304964A CN 112304964 A CN112304964 A CN 112304964A CN 202011378565 A CN202011378565 A CN 202011378565A CN 112304964 A CN112304964 A CN 112304964A
- Authority
- CN
- China
- Prior art keywords
- picture
- template
- value
- rectangular area
- fuse box
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/89—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles
- G01N21/892—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles characterised by the flaw, defect or object feature examined
- G01N21/898—Irregularities in textured or patterned surfaces, e.g. textiles, wood
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/001—Industrial image inspection using an image reference approach
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N2021/845—Objects on a conveyor
- G01N2021/8455—Objects on a conveyor and using position detectors
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8854—Grading and classifying of flaws
- G01N2021/888—Marking defects
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8883—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges involving the calculation of gauges, generating models
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
Abstract
The invention discloses a visual detection system for a fuse box, which comprises an industrial camera arranged in a camera bellows, a single plate to be detected penetrates through the camera bellows through a transmission belt arranged at the bottom of the camera bellows, and a camera module arranged in the camera bellows is triggered to take a picture through an infrared photoelectric switch; during detection, the single plate moves along with the conveyor belt to block infrared rays emitted by the infrared photoelectric switch perpendicular to the moving direction of the single plate, and the industrial camera is triggered to shoot pictures; the industrial camera sends the collected pictures to a computer; visual detection software is arranged in the computer; visual detection software of the computer carries out picture identification comparison on the shot picture and a preset template picture serving as a standard; if the comparison result meets the preset error range, the veneer visual detection is qualified, and the next fuse box veneer is submitted for inspection; if the error is not qualified, a warning is given and the content exceeding the error range is displayed on a computer display device.
Description
Technical Field
The invention relates to the field of computer vision identification, in particular to a fuse box vision detection system.
Background
Computer vision recognition is often viewed as a branch of artificial intelligence and computer science, and is now beginning to be deployed and used in large quantities in various industries. Moreover, these recognitions require specific lighting, background, and target pose requirements in a particular environment. Therefore, in practice, it is not easy to realize the desired detection target due to the limitations of complexity, external environment, cost, and the like of the detected object. By increasing the performance of the camera, enabling the camera to move physically uniformly, applying a striking color or shape to the detected target and requiring an ideal condition of the surrounding environment, the difficulty of visual identification can be reduced, and the accuracy can be improved. In the process of mass repetitive industrial production, the machine vision detection method can greatly improve the production efficiency and the automation degree.
Taking the production of fuse boxes as an example, since the detection of the fuse capacity cannot be judged by the electrical property, it is difficult to judge by the test of the electrical property for a possible error in the installation. The device layout of each kind of fuse box single plate is unchanged, a plurality of devices with different color blocks are distributed on the fuse box single plate, the devices with different color blocks represent different kinds of electronic elements, and therefore the fuse box single plate cannot be confused. Therefore, the types of the single plates of the fuse box are various, and the number of the devices to be inserted into one single plate is large, so that the difficulty of manual detection is caused, namely, the accuracy and the speed cannot be ensured.
In the existing method, a visual recognition system is adopted to detect the color of the fuse, the application number of the prior art is 201811285912, and the invention name is: the visual identification and detection system for the automobile fuse box comprises an image identification system, an industrial personal computer and a PLC, wherein the image identification system of an intelligent camera is used for identifying fuses in a picture and sending the picture to the industrial personal computer and the PLC; and the PLC judges whether the tested picture is qualified. In the method, only one frame is mentioned for the process of visual identification, and no more details exist, so that the method is not beneficial to popularization and use.
The visual detection system provided by the invention can be used for a production line of fuse box products. The method comprises the steps of carrying out image recognition on a single board to be detected by using a camera, extracting color layout information of various devices on the single board, finding a device installation error part of a worker on a production line by comparing standard templates, giving an alarm, assisting in finding a fault part and improving the product percent of pass.
Disclosure of Invention
1. The technical problem to be solved is as follows:
aiming at the technical problem, the invention provides a fuse box visual detection system, which is implemented in the scene of product visual identification on an industrial production line, and utilizes a computer visual identification technology to visually detect each key area of a fuse box product so as to find devices which do not meet the standard; by the visual detection system, abnormal devices on product installation can be found, errors can be found and corrected in time, and the product assembly accuracy is improved.
2. The technical scheme is as follows:
a fuse box visual inspection system characterized in that: comprises an industrial camera, a camera bellows, a conveyor belt and a computer; the industrial camera is arranged at the top of the camera bellows; a plurality of shooting light sources are arranged in the camera bellows; the bottom of the camera bellows is provided with an inlet and an outlet which are opposite to each other and enter and exit the fuse box; limiting plates are respectively arranged between the inlet and the outlet along the movement direction of the single plate of the fuse box; the limiting plate extends out of the camera bellows at the inlet, so that the initial position of the single plate of the fuse box is convenient to position when the single plate of the fuse box is placed; the camera bellows is fixed on the conveyer belt; during detection, the detected fuse box single plate is placed on a conveyor belt at an inlet and moves with the conveyor belt to pass through a dark box; a transmitting device and a receiving device of the infrared photoelectric switch are arranged in the dark box in the direction vertical to the moving direction of the single plate of the fuse box; when the infrared photoelectric switch is triggered, namely the moment that the single plate of the fuse box shields the infrared of the infrared photoelectric switch in the camera bellows, the industrial camera is triggered to take a picture; the industrial camera sends the collected pictures to a computer; visual detection software is arranged in the computer; visual detection software of the computer carries out picture identification comparison on the shot picture and a preset template picture serving as a standard; if the comparison result meets the preset error range, the veneer visual detection is qualified, and the next fuse box veneer is submitted for inspection; if the error is not qualified, a warning is given and the content exceeding the error range is displayed on a computer display device.
Further, the visual detection software comprises template making and picture identification; the template manufacturing specifically comprises: putting a standard single board with a fixed model on the surface of a conveyor belt, entering a camera bellows, acquiring pictures by an industrial camera, sending the pictures to a computer, selecting rectangular areas by a detector from the areas corresponding to each part to be detected in the pictures by using a mouse frame, marking the rectangular areas one by one, and acquiring coordinate information corresponding to each rectangular area by software; storing the picture and the marked rectangular area information; and finally, naming and storing the template.
Further, the process of the visual inspection software for image recognition specifically includes:
s1 selecting a template; selecting a template with the same type as a single plate of the fuse box to be detected, acquiring each rectangular area marked on the template, and calculating the color characteristic value of each rectangular area by adopting an algorithm; the template making according to claim 2 is performed if there is no existing template.
S2 picture collection; starting a conveyor belt, scanning product one-dimensional code information of the code single plate, and placing the fuse box single plate to be detected at the inlet of the camera bellows; the conveying belt conveys the fuse box single plate to be detected to the interior of the camera bellows, after the infrared photoelectric switch is triggered, the industrial camera collects the picture of the fuse box single plate entering the camera bellows, the picture is conveyed to the computer, and the visual detection system carries out picture identification.
S3 picture processing and recognition; the method comprises the following steps of comparing a shot picture received by a computer with a picture of a template, identifying an unqualified rectangular area and giving an alarm, and specifically comprises the following steps: calculating the color characteristic value of each corresponding rectangular area on the shot picture according to the position information of each rectangular area pre-stored in the template; comparing the color characteristic values of the rectangular areas at the same positions on the shot picture and the picture of the template to obtain the color distance value of the image of the rectangular area as the difference of the rectangular area; judging whether each rectangular area on the collected picture is consistent with the template; if the color moment value is smaller, the two comparison images are closer.
S4 the computer displays the picture of the template and the shot picture on the display device of the computer at the same time, and marks the corresponding rectangular area exceeding the preset color distance value in the collected picture; and simultaneously, independently displaying the corresponding rectangular area exceeding the preset color distance value for checking and correcting by detection personnel according to the displayed information.
Further, the picture processing and recognition further includes a picture offset correction process, and the picture offset correction process specifically includes: in the template manufacturing process, selecting an area with obvious characteristics and fixed position in a picture of a single plate of the fuse box as a positioning area, drawing a positioning rectangle by using a mouse and storing the positioning rectangle as template information; during the process of processing and identifying the picture, finding a rectangular area which is closest to or most matched with the positioning rectangle in the shot picture, and obtaining the deviation of the shot picture on the X axis and the Y axis compared with the picture of the template by comparing the difference value between the coordinate values of the area and the positioning area on the template; and performing overall offset correction on the shot picture on an X axis and a Y axis according to the offset information.
Further, when the picture is processed and identified, judging whether each rectangular area on the collected picture is consistent with the template or not, and adopting an improved method of color moment average value; the improvement method of the color moment average value specifically comprises the following steps: the color moment average value is obtained by comparing all rectangular area pictures with the pictures of the template, the improved value of the color moment average value is the color moment average value multiplied by a preset coefficient, and if the color moment value of each rectangular area is larger than the improved value of the average value, the system determines that the difference between the rectangular area of the shot picture and the template is too large, and an alarm is needed; if the color moment value of the rectangular area is smaller than the improved value of the average value, the system determines that the rectangular area is not greatly different from the template, and does not need to give an alarm; the improved value of the color matrix average value is used for judging the similarity of the shot picture as a whole relative to the template, and when the color matrix value of a certain rectangular area is larger than the improved value of the average value, the rectangular area belongs to a false sharp detection area, and the probability of a fault piece in the area is the highest.
Further, when the picture is processed and identified, judging whether each rectangular area on the collected picture is consistent with the template or not, and adopting an improved method of the maximum value and the minimum value of the color moment; the improvement method of the maximum value and the minimum value of the color moment specifically comprises the following steps: the method comprises the following steps that a detector presets a maximum value of a color moment and a minimum value of a color distance according to an actual running condition and an actual detection result; if the color moment value of the rectangular area in the shot picture is smaller than the minimum color distance value, no alarm is sent; and if the color moment value of the rectangular area is larger than the preset maximum color distance value, an alarm is given.
3. Has the advantages that:
(1) according to the invention, the transmission belt is adopted to transmit the single board to be detected, and the infrared photoelectric switch trigger mechanism is arranged to control the industrial camera to shoot pictures, so that the aim of identifying the pictures in motion can be realized.
(2) According to the invention, the template serving as the standard is preset and stored, and then the picture of the veneer to be detected is compared with the template, so that the production efficiency is effectively improved.
(3) When the template is arranged, the part to be checked in the picture is firstly drawn into a rectangle in the template by using a mouse and the position information and the picture information of each rectangle are stored, so that the comparison with the picture of the single board to be detected can be conveniently carried out by one rectangle, the rectangular area which does not accord with the similarity requirement in comparison can be marked, and the later-stage overhaul can be favorably carried out by a maintainer in a targeted manner. And can also be used to identify areas of relatively small area, such as small screw and nut sizes.
In conclusion, the vision detection system provided by the invention adopts a computer vision identification algorithm, and well solves the problem of finding the positions and the number of the inserted wrong devices through acquisition, identification and alarm, can find the problem single board earlier, and avoids the problem products from being delivered out of the warehouse.
Drawings
FIG. 1 is an apparatus front view of a fuse block visual inspection system of the present invention;
FIG. 2 is a top plan view of the apparatus of the fuse box visual inspection system of the present invention;
FIG. 3 is a flowchart of the visual inspection software for image recognition according to the present invention;
FIG. 4 is a flowchart of the creation of a template for the visual inspection software of the present invention;
FIG. 5 is a schematic diagram of the present invention.
Detailed Description
The present invention will be described in detail with reference to the accompanying drawings.
As shown in fig. 1 and 2, a fuse box visual inspection system is characterized in that: comprises an industrial camera 4, a camera bellows 1, a conveyor belt 6 and a computer; the industrial camera is arranged at the top of the camera bellows; a plurality of shooting light sources are arranged in the camera bellows; the bottom of the camera bellows is provided with an inlet 3 and an outlet 2 which are opposite to each other and enter and exit the fuse box; limiting plates 8 are respectively arranged between the inlet and the outlet along the movement direction of the fuse box single plate 7; the limiting plate extends out of the camera bellows at the inlet, so that the initial position of the single plate of the fuse box is convenient to position when the single plate 7 of the fuse box is placed; the camera bellows is fixed on the conveyer belt; during detection, the detected fuse box single plate is placed on the conveyor belt 6 at the inlet and moves with the conveyor belt to pass through the camera bellows; a transmitting device and a receiving device of the infrared photoelectric switch are arranged in the dark box in the direction vertical to the moving direction of the single plate of the fuse box; referring to fig. 5, when the infrared photoelectric switch 5 is triggered, that is, the instant when the single plate of the fuse box shields the infrared ray of the infrared photoelectric switch in the camera bellows, the industrial camera is triggered to take a picture; the industrial camera sends the collected pictures to a computer; visual detection software is arranged in the computer; visual detection software of the computer carries out picture identification comparison on the shot picture and a preset template picture serving as a standard; if the comparison result meets the preset error range, the veneer visual detection is qualified, and the next fuse box veneer is submitted for inspection; if the error is not qualified, a warning is given and the content exceeding the error range is displayed on a computer display device.
In the invention, the computer is internally provided with visual detection software; visual detection software of the computer identifies and compares the shot picture with a preset standard template picture; if the comparison result is in line with the expectation, the product is qualified, and the next submission is carried out; if the product is not qualified, a warning is given and the warning content is displayed on a computer display device, so that the fault piece can be found in an assisted manner, and the product qualification rate is improved.
Further, the visual detection software comprises template making and picture identification; the template manufacturing specifically includes, as shown in fig. 4: putting a standard single board with a fixed model on the surface of a conveyor belt, entering a camera bellows, acquiring pictures by an industrial camera, sending the pictures to a computer, selecting rectangular areas by a detector from the areas corresponding to each part to be detected in the pictures by using a mouse frame, marking the rectangular areas one by one, and acquiring coordinate information corresponding to each rectangular area by software; storing the picture and the marked rectangular area information; and finally, naming and storing the template.
In the invention, by collecting the picture information of the single plate of the standard fuse box and marking out the devices to be compared, the method is beneficial to the targeted comparison of the next step.
Further, as shown in fig. 3, the process of the visual inspection software for image recognition specifically includes:
s1 selecting a template; selecting a template with the same type as a single plate of the fuse box to be detected, acquiring each rectangular area marked on the template, and calculating the color characteristic value of each rectangular area by adopting an algorithm; the template making according to claim 2 is performed if there is no existing template.
S2 picture collection; starting a conveyor belt, scanning product one-dimensional code information of the code single plate, and placing the fuse box single plate to be detected at the inlet of the camera bellows; the conveying belt conveys the fuse box single plate to be detected to the interior of the camera bellows, after the infrared photoelectric switch is triggered, the industrial camera collects the picture of the fuse box single plate entering the camera bellows, the picture is conveyed to the computer, and the visual detection system carries out picture identification.
S3 picture processing and recognition; the method comprises the following steps of comparing a shot picture received by a computer with a picture of a template, identifying an unqualified rectangular area and giving an alarm, and specifically comprises the following steps: calculating the color characteristic value of each corresponding rectangular area on the shot picture according to the position information of each rectangular area pre-stored in the template; comparing the color characteristic values of the rectangular areas at the same positions on the shot picture and the picture of the template to obtain the color distance value of the image of the rectangular area as the difference of the rectangular area; judging whether each rectangular area on the collected picture is consistent with the template; if the color moment value is smaller, the two comparison images are closer.
S4 the computer displays the picture of the template and the shot picture on the display device of the computer at the same time, and marks the corresponding rectangular area exceeding the preset color distance value in the collected picture; and simultaneously, independently displaying the corresponding rectangular area exceeding the preset color distance value for checking and correcting by detection personnel according to the displayed information.
In the above steps, the color distance of the rectangular area image, that is, the difference between the two rectangular pictures, is obtained by comparing the color feature values of the same position area on the shot picture and the picture of the template. Therefore, the color moment values of the shot picture compared with the color moment values of all the areas of the template can be obtained, and whether all the detection areas on the collected picture are consistent with the template or not can be preliminarily judged according to the color moment values. The smaller the color moment value, the closer the two comparison images are identified, and the higher the similarity. When two identical images are aligned, the color moment value is zero. The color characteristic values of the rectangular areas marked in the template are compared, so that the size of the image compared at a time is reduced, and the accuracy of the comparison result of the device is improved.
Further, the picture processing and recognition further includes a picture offset correction process, and the picture offset correction process specifically includes: in the template manufacturing process, selecting an area with obvious characteristics and fixed position in a picture of a single plate of the fuse box as a positioning area, drawing a positioning rectangle by using a mouse and storing the positioning rectangle as template information; during the process of processing and identifying the picture, finding a rectangular area which is closest to or most matched with the positioning rectangle in the shot picture, and obtaining the deviation of the shot picture on the X axis and the Y axis compared with the picture of the template by comparing the difference value between the coordinate values of the area and the positioning area on the template; and performing overall offset correction on the shot picture on an X axis and a Y axis according to the offset information. In practical applications, the coordinates are by default with the upper left corner of the picture as the origin.
In the step, by comparing the coordinate values of the positioning area in the shot picture and the positioning area on the template, the offset of the collected picture on the X axis and the Y axis can be estimated, and the offset correction is performed on the collected picture according to the offset information, so that the accuracy of other areas in calculating the color characteristic value can be improved, and the comparison effect is enhanced.
Further, when the picture is processed and identified, judging whether each rectangular area on the collected picture is consistent with the template or not, and adopting an improved method of color moment average value; the improvement method of the color moment average value specifically comprises the following steps: the color moment average value is obtained by comparing all rectangular area pictures of the collected pictures with the pictures of the template, the improved value of the color moment average value is the color moment average value multiplied by a preset coefficient, and if the color moment value of each rectangular area is larger than the improved value of the average value, the system determines that the difference between the rectangular area of the shot picture and the template is too large, and an alarm is needed; if the color moment value of the rectangular area is smaller than the improved value of the average value, the system determines that the rectangular area is not greatly different from the template, and does not need to give an alarm; the improved value of the color matrix average value is used for judging the similarity of the shot picture as a whole relative to the template, and when the color matrix value of a certain rectangular area is larger than the improved value of the average value, the rectangular area belongs to a false sharp detection area, and the probability of a fault piece in the area is the highest.
The improved value of the color moment average value adopted in the step is represented by the similarity of the collected picture to the template on the whole, and when the color moment value of a certain area is larger than the average value multiplied by a coefficient, the area belongs to a false sharp detection area, and the probability of a fault part in the area is the largest.
Further, when the picture is processed and identified, judging whether each rectangular area on the collected picture is consistent with the template or not, and adopting an improved method of the maximum value and the minimum value of the color moment; the improvement method of the maximum value and the minimum value of the color moment specifically comprises the following steps: the method comprises the following steps that a detector presets a maximum value of a color moment and a minimum value of a color distance according to an actual running condition and an actual detection result; if the color moment value of the rectangular area in the shot picture is smaller than the minimum color distance value, no alarm is sent; and if the color moment value of the rectangular area is larger than the preset maximum color distance value, an alarm is given.
In the step, when judging whether each area is consistent with the template, introducing an improved method of the maximum value and the minimum value of the color moment: the maximum value and the minimum value of the color moment are two parameters set according to the actual running condition and the actual detection result of the system. For example, setting a minimum color moment value 40 and a maximum color moment value 50, if the color moment value of each area is smaller than the minimum value 40, no alarm is given, even if the color moment value is larger than the improved value of the color moment average value obtained in the previous step; if its colour moment value is greater than the maximum value 50, it will alarm, even if its colour moment value is less than the improved value of the previously obtained colour moment average value.
The maximum value and the minimum value of the color moment represent the image acquisition and visual recognition capability of the whole detection recognition system. When the color moment value of a certain area is larger than the set maximum value, the probability of the occurrence of a fault element is considered to be high, because the color moments of a plurality of areas on the collected image are likely to be large, the color moment average value is large due to the occurrence of the fault element in the plurality of areas, and the color moment average value has no reference meaning at the moment, the maximum value of the color moment is introduced.
For the collected picture with higher similarity to the template, the situation that the average value of the color moment is very small can occur during comparison, the color moment of an individual false sharp area is larger than the improved value of the average value at the moment, if the method of only comparing the improved value of the average value is adopted, the abnormal area is considered, however, in practice, the area does not form an alarm and does not belong to a fault piece, the minimum value of the color moment value is required to be used, and the situation of false alarm is filtered out. I.e. the colour moment value of the area is less than the minimum value, no alarm is given, even if the colour moment value of the area is greater than the improvement of the average value. Therefore, in actual work, two methods are usually used in a superposition manner to judge whether to alarm or not.
Although the present invention has been described with reference to the preferred embodiments, it should be understood that various changes and modifications can be made therein by those skilled in the art without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (6)
1. A fuse box visual inspection system characterized in that: comprises an industrial camera, a camera bellows, a conveyor belt and a computer; the industrial camera is arranged at the top of the camera bellows; a plurality of shooting light sources are arranged in the camera bellows; the bottom of the camera bellows is provided with an inlet and an outlet which are opposite to each other and enter and exit the fuse box; limiting plates are respectively arranged between the inlet and the outlet along the movement direction of the single plate of the fuse box; the limiting plate extends out of the camera bellows at the inlet, so that the initial position of the single plate of the fuse box is convenient to position when the single plate of the fuse box is placed; the camera bellows is fixed on the conveyer belt; during detection, the detected fuse box single plate is placed on a conveyor belt at an inlet and moves with the conveyor belt to pass through a dark box; a transmitting device and a receiving device of the infrared photoelectric switch are arranged in the dark box in the direction vertical to the moving direction of the single plate of the fuse box; when the infrared photoelectric switch is triggered, namely the moment that the single plate of the fuse box shields the infrared of the infrared photoelectric switch in the camera bellows, the industrial camera is triggered to take a picture; the industrial camera sends the collected pictures to a computer; visual detection software is arranged in the computer; visual detection software of the computer carries out picture identification comparison on the shot picture and a preset template picture serving as a standard; if the comparison result meets the preset error range, the veneer visual detection is qualified, and the next fuse box veneer is submitted for inspection; if the error is not qualified, a warning is given and the content exceeding the error range is displayed on a computer display device.
2. The visual inspection system of claim 1, wherein the visual inspection software includes template making and picture recognition; the template manufacturing specifically comprises: putting a standard single board with a fixed model on the surface of a conveyor belt, entering a camera bellows, acquiring pictures by an industrial camera, sending the pictures to a computer, selecting rectangular areas by a detector from the areas corresponding to each part to be detected in the pictures by using a mouse frame, marking the rectangular areas one by one, and acquiring coordinate information corresponding to each rectangular area by software; storing the picture and the marked rectangular area information; and finally, naming and storing the template.
3. A visual fuse box inspection system as recited in claim 2, wherein: the process of the visual inspection software for picture identification specifically comprises the following steps:
s1 selecting a template; selecting a template with the same type as a single plate of the fuse box to be detected, acquiring each rectangular area marked on the template, and calculating the color characteristic value of each rectangular area by adopting an algorithm; if there is no existing template, performing the template making according to claim 2;
s2 picture collection; starting a conveyor belt, scanning product one-dimensional code information of the code single plate, and placing the fuse box single plate to be detected at the inlet of the camera bellows; the method comprises the following steps that a conveyor belt conveys a fuse box single plate to be detected to the interior of a camera bellows, an industrial camera collects a picture of the fuse box single plate entering the camera bellows after an infrared photoelectric switch is triggered, the picture is conveyed to a computer, and a visual detection system identifies the picture;
s3 picture processing and recognition; the method comprises the following steps of comparing a shot picture received by a computer with a picture of a template, identifying an unqualified rectangular area and giving an alarm, and specifically comprises the following steps: calculating the color characteristic value of each corresponding rectangular area on the shot picture according to the position information of each rectangular area pre-stored in the template; comparing the color characteristic values of the rectangular areas at the same positions on the shot picture and the picture of the template to obtain the color distance value of the image of the rectangular area as the difference of the rectangular area; judging whether each rectangular area on the collected picture is consistent with the template; if the color moment value is smaller, the two comparison images are closer;
s4 the computer displays the picture of the template and the shot picture on the display device of the computer at the same time, and marks the corresponding rectangular area exceeding the preset color distance value in the collected picture; and simultaneously, independently displaying the corresponding rectangular area exceeding the preset color distance value for checking and correcting by detection personnel according to the displayed information.
4. A visual fuse box inspection system as set forth in claim 3 wherein: the picture processing and recognition further comprises a picture offset correction process, wherein the picture offset correction process specifically comprises the following steps: in the template manufacturing process, selecting an area with obvious characteristics and fixed position in a picture of a single plate of the fuse box as a positioning area, drawing a positioning rectangle by using a mouse and storing the positioning rectangle as template information; during the process of processing and identifying the picture, finding a rectangular area which is closest to or most matched with the positioning rectangle in the shot picture, and obtaining the deviation of the shot picture on the X axis and the Y axis compared with the picture of the template by comparing the difference value between the coordinate values of the area and the positioning area on the template; and performing overall offset correction on the shot picture on an X axis and a Y axis according to the offset information.
5. A visual fuse box inspection system as set forth in claim 3 wherein: when the picture is processed and identified, judging whether each rectangular area on the collected picture is consistent with the template or not, and adopting an improved method of color moment average value; the improvement method of the color moment average value specifically comprises the following steps: the color moment average value is obtained by comparing all rectangular area pictures with the pictures of the template, the improved value of the color moment average value is the color moment average value multiplied by a preset coefficient, and if the color moment value of each rectangular area is larger than the improved value of the average value, the system determines that the difference between the rectangular area of the shot picture and the template is too large, and an alarm is needed; if the color moment value of the rectangular area is smaller than the improved value of the average value, the system determines that the rectangular area is not greatly different from the template, and does not need to give an alarm; the improved value of the color matrix average value is used for judging the similarity of the shot picture as a whole relative to the template, and when the color matrix value of a certain rectangular area is larger than the improved value of the average value, the rectangular area belongs to a false sharp detection area, and the probability of a fault piece in the area is the highest.
6. A visual fuse box inspection system as set forth in claim 3 wherein: when the picture is processed and identified, judging whether each rectangular area on the collected picture is consistent with the template or not, and adopting an improved method of the maximum value and the minimum value of the color moment; the improvement method of the maximum value and the minimum value of the color moment specifically comprises the following steps: the method comprises the following steps that a detector presets a maximum value of a color moment and a minimum value of a color distance according to an actual running condition and an actual detection result; if the color moment value of the rectangular area in the shot picture is smaller than the minimum color distance value, no alarm is sent; and if the color moment value of the rectangular area is larger than the preset maximum color distance value, an alarm is given.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011378565.5A CN112304964A (en) | 2020-12-01 | 2020-12-01 | Fuse box vision detection system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011378565.5A CN112304964A (en) | 2020-12-01 | 2020-12-01 | Fuse box vision detection system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112304964A true CN112304964A (en) | 2021-02-02 |
Family
ID=74487398
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011378565.5A Pending CN112304964A (en) | 2020-12-01 | 2020-12-01 | Fuse box vision detection system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112304964A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113624145A (en) * | 2021-08-09 | 2021-11-09 | 吉林大学 | Visual realization method for measuring assembly height and inclination degree of device in automobile fuse box |
| CN114236885A (en) * | 2021-11-10 | 2022-03-25 | 云南电网有限责任公司 | Visual detection system and method for electric energy meter liquid crystal display machine |
| CN116168021A (en) * | 2023-04-21 | 2023-05-26 | 中江立江电子有限公司 | Fault part identification system and method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1746667A (en) * | 2004-09-06 | 2006-03-15 | 欧姆龙株式会社 | Substrate inspection method and apparatus |
| CN1797426A (en) * | 2004-12-27 | 2006-07-05 | 欧姆龙株式会社 | Image processing method, substrate inspection method, substrate inspection apparatus and method of generating substrate inspection data |
| CN1865950A (en) * | 2005-04-25 | 2006-11-22 | 王�琦 | Automatic aligning method for printed circuit board |
| CN105158268A (en) * | 2015-09-21 | 2015-12-16 | 武汉理工大学 | Intelligent online detection method, system and device for defects of fine-blanked parts |
| CN109521015A (en) * | 2018-10-31 | 2019-03-26 | 上海沪工汽车电器有限公司 | Automobile fuse box visual identity detection system |
-
2020
- 2020-12-01 CN CN202011378565.5A patent/CN112304964A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1746667A (en) * | 2004-09-06 | 2006-03-15 | 欧姆龙株式会社 | Substrate inspection method and apparatus |
| CN1797426A (en) * | 2004-12-27 | 2006-07-05 | 欧姆龙株式会社 | Image processing method, substrate inspection method, substrate inspection apparatus and method of generating substrate inspection data |
| CN1865950A (en) * | 2005-04-25 | 2006-11-22 | 王�琦 | Automatic aligning method for printed circuit board |
| CN105158268A (en) * | 2015-09-21 | 2015-12-16 | 武汉理工大学 | Intelligent online detection method, system and device for defects of fine-blanked parts |
| CN109521015A (en) * | 2018-10-31 | 2019-03-26 | 上海沪工汽车电器有限公司 | Automobile fuse box visual identity detection system |
Non-Patent Citations (2)
| Title |
|---|
| 郑继刚等著: "《基于MATLAB的数字图像处理研究》", 31 December 2010, 云南大学出版社 * |
| 韩九强等著: "《数字图像处理基于XAVIS组态软件》", 30 September 2018, 西安交通大学出版社 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113624145A (en) * | 2021-08-09 | 2021-11-09 | 吉林大学 | Visual realization method for measuring assembly height and inclination degree of device in automobile fuse box |
| CN113624145B (en) * | 2021-08-09 | 2022-04-26 | 吉林大学 | Visual realization method for measuring assembly height and inclination degree of device in automobile fuse box |
| CN114236885A (en) * | 2021-11-10 | 2022-03-25 | 云南电网有限责任公司 | Visual detection system and method for electric energy meter liquid crystal display machine |
| CN114236885B (en) * | 2021-11-10 | 2023-09-12 | 云南电网有限责任公司 | Electric energy meter liquid crystal display machine vision detection system and method |
| CN116168021A (en) * | 2023-04-21 | 2023-05-26 | 中江立江电子有限公司 | Fault part identification system and method |
| CN116168021B (en) * | 2023-04-21 | 2023-08-29 | 中江立江电子有限公司 | Fault part identification system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112304964A (en) | Fuse box vision detection system | |
| US11213860B2 (en) | Automatic magnetic core sorting system based on machine vision | |
| CN101566582B (en) | Medicine bottle label information online detection system in powder injection production based on mechanical vision | |
| US10438340B2 (en) | Automatic optical inspection system and operating method thereof | |
| CN112304965A (en) | Multi-scene fuse box visual detection system | |
| CN107270841A (en) | A kind of packaging bag coding defects detection mechanism | |
| CN111551559A (en) | LCD (liquid Crystal display) liquid crystal screen defect detection method based on multi-view vision system | |
| CN114324398A (en) | Intelligent visual detection system and detection method for medical and edible packaging label | |
| CN114518526A (en) | Automatic testing machine control system suitable for PCB board ICT | |
| CN116993725B (en) | Intelligent patch information processing system of flexible circuit board | |
| CN113916285A (en) | Visual detection device and detection process for washing machine roller fastening screw | |
| CN114359155A (en) | Film laminating method and system | |
| CN111275671A (en) | robot-3D technology-matched visual identification method | |
| CN116718600A (en) | Detection device for intelligent detection of wire harness product | |
| CN115601345A (en) | Machine vision-based large-size punching plate hole group leakage-proof detection method and device | |
| CN109158325A (en) | The device for visual identification of IC punching system kind and recognition methods and application | |
| CN110856847A (en) | Capacitance character detection method and device based on intelligent vision | |
| CN117197112B (en) | Industrial screen printing defect detection method and system based on sample comparison | |
| CN106023174B (en) | A kind of novel detection money thousand sheets packet is several devices and methods therefors | |
| CN117358615B (en) | Automatic code-spraying printing defect detection method and system | |
| CN116087217B (en) | Industrial assembly line dynamic quality detection module and method based on machine vision | |
| CN114783000B (en) | Method and device for detecting dressing standard of worker in bright kitchen range scene | |
| Cheng | Design of Visual Search and Positioning System Based on Labview+ PLC | |
| CN117115114B (en) | YOLO-based power electronic device production process target identification and defect detection method | |
| CN114708234A (en) | Method and device for identifying number of detonators on automatic bayonet coding all-in-one machine |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210202 |