CN108545265A - A kind of tablet capsule counting device and its method based on machine vision - Google Patents

Abstract

The invention discloses a tablet and capsule counting device and method based on machine vision, including a CCD camera mechanism; the invention adopts a CCD camera mechanism that can adjust the height position up and down and the horizontal position back and forth, and adjust the bulk material through the gradient adjustment mechanism. The adjustment function of the inclination angle of the plate, and the vibration effect of the vibration mechanism on the bulk material plate, make a number of tablet capsules arranged and accommodated in the bottom of the W-shaped groove and form a row of tablet capsules, which greatly improves To improve the quality of the collected true-color images of tablets and capsules; use MATLAB tools to process the images of true-color images of tablets and capsules through image processing methods, and count the number of tablets and capsules accurately according to the image processing results, which greatly improves the count of tablets and capsules The accuracy and work efficiency; and the overall design of the device is reasonable, compact, make full use of the limited plane and space position, low cost, suitable for popularization and application.

Description

A device and method for counting tablets and capsules based on machine vision

technical field

The invention relates to a counting device, in particular to a machine vision-based tablet and capsule counting device and method thereof.

Background technique

Aiming at the pharmaceutical manufacturers, in the process of packaging tablets and capsules, it is necessary to manually pick up the tablets and capsules with a capsule shovel and manually count the number of the scooped tablets and capsules, and then complete the boxing and packaging of the tablets and capsules, which makes the counting inaccurate and greatly Increased labor intensity of workers, greatly reducing work efficiency is not high.

For hospitals, pharmacies and other places where the pharmaceutical industry is used, medical personnel need to reclassify and repackage the tablets and capsules of various drugs according to a certain number according to requirements. Due to the large variety of drugs and the medical personnel Extremely limited, if the tablets and capsules are poured out one by one and manually counted, and then classified, not only will the medical staff have a huge workload and extremely low work efficiency, but also the method of manually counting the tablets and capsules is extremely easy to count the wrong number, Bring unnecessary losses to hospitals, pharmacies or users who buy tablet capsules.

Contents of the invention

The technical problem solved by the present invention is to overcome the defects of the prior art and provide a machine vision-based tablet capsule counting device that adopts a CCD camera mechanism that can adjust the height position up and down and the horizontal position can be adjusted back and forth to collect the true color images of the tablet capsules .

In order to solve the above technical problems, the present invention is solved through the following technical solutions:

A tablet and capsule counting device based on machine vision, comprising a chassis, a bulk material mechanism fixed in the chassis and used to hold tablets and capsules, a CCD camera mechanism fixed above the chassis, a host computer module connected to the CCD camera mechanism, and The display screen connected to the upper computer module, the top of the chassis is provided with a window facing the upper CCD camera mechanism, the bulk material mechanism includes a bulk material plate that can be tilted and fixed under the window through an inclination adjustment mechanism, and the bulk material plate is equipped with Several W-shaped grooves, several tablets and capsules are arranged and accommodated in the bottom of the W-shaped grooves to form a row of regularly arranged rows of tablets and capsules; the CCD camera mechanism includes adjustable height positions up and down or front and rear adjustment levels The industrial CCD camera at the position, the industrial CCD camera is fixed on the top of the chassis and is directly facing the window below; the row of regularly arranged tablet capsules formed in the bottom of the W-shaped groove is photographed by the industrial CCD camera And get the original true-color image, the industrial CCD camera then transmits the collected original true-color image to the host computer module, the host computer module uses MATLAB tools to process the original true-color image through image processing methods, according to the image Processing Results Count the number of tablets and capsules in the original true-color image to obtain the number of tablets and capsules in the bottom of the W-shaped groove, and the number of tablets and capsules is displayed on the display screen. Among them, the inclination state of the bulk material plate is low at the front and high at the rear, the front side of the bulk material plate is fixed in the chassis, and the rear side of the bulk material plate is driven up or down by the inclination adjustment mechanism; the upper computer module is set in the upper computer, The display screen can be installed on the chassis or on the upper computer; it adopts the combined structure setting of the CCD camera mechanism that can adjust the height position up and down and the horizontal position can be adjusted forward and backward. Through the position adjustment function of the CCD camera mechanism and the tilt adjustment mechanism The adjustment function of the inclination angle of the bulk material plate, and through the bulk material mechanism, several tablet capsules are arranged and accommodated in the bottom of the W-shaped groove to form a row of tablet capsule queues, which greatly improves the collection of tablet capsules. The quality of the true color image; using MATLAB tools, the image processing method is used to image the true color image of the tablet and capsule, and the accurate number of the tablet and capsule is counted according to the image processing result, which greatly improves the accuracy and work of the tablet and capsule count. efficiency; and the overall design of the device is reasonable, the structure is compact, the limited plane and space are fully utilized, and the cost is low.

Preferably, the angle adjustment range of the inclination adjustment mechanism for the inclination angle between the bulk material plate and the horizontal plane is 15-30 degrees. When several tablets and capsules are put into the bulk material plate, the bulk material plate in a 15-30 degree inclined state enables the tablets and capsules to slide automatically along the inclined wall of the bulk material plate into the W In the groove bottom of the type groove, it not only avoids the stacking of tablets and capsules, but also reduces the number of bonding points between tablets and capsules, which reduces the difficulty of subsequent counting of tablets and capsules through image processing methods and improves work efficiency. , and can ensure the accuracy of the number of tablets and capsules in the original true-color image captured by the follow-up industrial CCD camera, so that the accuracy of the number of tablets and capsules can be obtained by processing and counting the original true-color image through the image processing method If the included angle of inclination is less than 15-30 degrees, the slope of the inclined wall of the bulk material plate is not high enough, making it difficult for the tablet capsule to automatically follow the bulk material plate under the combined action of its own weight and inertia The inclined wall surface slides into the groove bottom of the W-shaped groove, making it difficult to accurately and effectively send several tablet capsules to be counted into the groove bottom of the W-shaped groove of the bulk material plate, thereby reducing work efficiency and Affects the accuracy of the number of tablets and capsules in the original true-color image captured by the industrial CCD camera; if the angle of inclination is higher than 15-30 degrees, it may cause the tablets and capsules that slide into the bottom of the W-shaped groove to reappear. Falling out of the bottom of the W-shaped groove makes it difficult to accurately and effectively send several tablet capsules to be counted into the bottom of the W-shaped groove of the bulk material plate, thereby reducing work efficiency and affecting the subsequent industrial CCD The accuracy of the number of pills and capsules in the raw true-color image captured by the camera.

Preferably, the inclination adjustment mechanism includes a push rod, a rotatable lead screw fixed at the bottom of the push rod and driving the push rod to move up and down, and the top of the push rod is fixed at the bottom of the rear side of the bulk material plate. The screw is rotated by external force, and the screw drives the ejector rod to move up or down at the same time, and the ejector rod moving up or down simultaneously drives the rear side of the bulk material plate to rise or fall, so that the bulk material plate is in a state of low front and high rear and the bulk material The inclination angle between the plate and the horizontal plane is 15-30 degrees, that is, when the angle between the bulk material plate and the horizontal plane is lower than 15-30 degrees, the screw is rotated by external force to drive the ejector rod to move upwards and make the The rear side of the bulk material plate rises relative to the front side until the angle reaches 15-30 degrees; when the angle between the bulk material plate and the horizontal plane is higher than 15-30 degrees, the screw is driven downward by external force Move and make the rear side of the bulk plate drop relative to the front side until the angle reaches 15-30 degrees.

Preferably, it also includes a bar-shaped light source fixed on both sides of the bulk material plate, and two opposite bar-shaped light sources simultaneously emit light to the bulk material plate and illuminate a row formed in the bottom of the W-shaped groove. The regularly arranged queue of tablets and capsules improves the image quality and imaging efficiency of the original true-color images collected by the industrial CCD camera for capturing the tablets and capsules.

As preferably, it also includes a vibration mechanism for vibrating the bulk material plate. The vibration mechanism includes a vibration shaft and a motor that drives the vibration shaft to generate vibration. The top of the vibration shaft is fixed at the center of the bottom of the bulk material plate, and the motor is fixed on the vibration shaft. bottom end. Among them, the motor is supported in the chassis; the vibration mechanism formed by the combination of the motor and the vibration shaft vibrates the bulk material plate in the inclined state, further enabling the tablets and capsules to automatically follow the bulk material under the combined action of self-weight and inertia. The inclined wall of the plate slides into the groove bottom of the W-shaped groove, which further reduces the possibility of stacking between the tablet capsules, further reduces the possibility of increasing the number of adhesion points between the tablet capsules, and further reduces the possibility of subsequent passage. It is difficult for the image processing method to count the tablets and capsules; if there is no vibration mechanism to vibrate the bulk plate, it is easy for some tablets and capsules to be attached to the groove wall of the W-shaped groove and collide with the bottom of the W-shaped groove. The tablets and capsules stick together, so that any two of at least three tablets and capsules are stuck together, and there are three sticking points in three tablets and capsules, so that the image segmentation in the subsequent image processing process needs to deal with more nodes. It greatly increases the difficulty of subsequent counting of tablets and capsules through image processing methods, and also causes large deviations or errors in the statistical counting of tablets and capsules after image segmentation, resulting in inaccurate counts and reduced processing efficiency; or Stacking other tablet capsules on top of the tablet capsules in the groove bottom of the W-shaped groove also greatly increases the difficulty and counting deviation or error of subsequent tablet capsule counting by image processing methods.

As preferably, the CCD camera mechanism also includes a support bar fixed on the rear side of the cabinet, a connecting block fixed on the support bar that can adjust the height position up and down or the horizontal position can be adjusted back and forth, fixedly connected on the connecting block and facing the direction of the window. The protruding camera clip, the upper and lower fine-tuning knobs and the front and rear fine-tuning knobs that are rotatably fixed between the connecting block and the support rod through the threaded hole, the industrial CCD camera is fixed in the camera clip through the fastening knob; Move the fine-tuning knob up and down, so that the connecting block moves up and down along the support rod, and at the same time drive the camera clip and industrial CCD camera to move up and down; turn the front and rear fine-tuning knobs through the threaded hole, so that the connecting block moves forward and backward along the supporting rod, and at the same time drive the camera clip and industrial CCD The camera moves back and forth. The combined setting of the upper and lower fine-tuning knobs and the front and rear fine-tuning knobs realizes the adjustment of the height position and the front and rear position of the industrial CCD camera, making it possible to form a row of regular rows of tablets and capsules in the bottom of the W-shaped groove. The vertical and horizontal distances of the industrial CCD camera are used to adjust the camera position of the industrial CCD camera, so that the industrial CCD camera can capture the original true-color image with the highest definition. The combined structural setting of the support rod, the connecting block, and the camera clip improves the reliability and stability of the industrial CCD camera in the vertical adjustment of the height position, the front and rear adjustment of the horizontal position, and the camera work process.

Preferably, one baffle plate is fixed on the front and rear sides of the bulk material plate, which prevents the tablet capsules from slipping or falling out of the bulk material plate.

Preferably, the inclination adjustment mechanism further includes a rotatable hand handle or electric handle fixed on the lead screw and driving the lead screw to rotate. The hand handle is convenient for manual operation; the electric handle realizes automatic operation and saves labor.

A method for counting tablets and capsules based on machine vision, comprising the steps of:

Step Ⅰ: Rotate the lead screw through the hand handle first, the lead screw drives the ejector rod to move up or down at the same time, and the ejector rod moving up or down simultaneously drives the rear side of the bulk material plate to rise or fall, so that the bulk material plate is forward Low back high state and the inclination angle between the bulk material plate and the horizontal plane is 15-30 degrees; then put several tablets and capsules to be counted into the bulk material plate, and then start the motor, the motor drives the vibration shaft to the bulk material plate A vibration effect is generated, and several tablet capsules slide into the W-shaped groove under the vibration effect until a row of tablet-capsule queues are formed in the groove bottom of each W-shaped groove;

Step Ⅱ: Then turn on two oppositely arranged bar-shaped light sources at the same time, and the two oppositely arranged bar-shaped light sources emit light to the bulk material plate at the same time and illuminate the rows of tablets formed in the bottom of the W-shaped groove capsule queue;

Step Ⅲ: Restart the industrial CCD camera, use the industrial CCD camera to take pictures of the illuminated rows of pills and capsules and get the original true-color image, and then the industrial CCD camera will transmit the collected original true-color image For the host computer module, the host computer module uses MATLAB tools to process the original true-color image through image processing methods, and counts several tablets and capsules in the original true-color image according to the image processing results to obtain W-shaped grooves The number of tablets and capsules in the bottom of the tank is displayed on the display screen.

As preferably, the image processing method in step III includes the following steps:

Step 1: Use MATLAB tools to read the original true-color image;

Step 2: Image preprocessing: including the following specific steps:

A. Image grayscale processing: According to the components of R, G, and B of each pixel in the original true-color image, the weight of each pixel is calculated through the Gray(i,j) function with R, G, and B as weights. Grayscale, so as to convert the original true color image into a grayscale image, where Gray(i,j)=0.299×R(i,j)+0.587×G(i,j)+0.114×B(i, j), Gray(i,j) is the gray level of the grayscale image at i,j, R(i,j) is the red brightness value of the original true color image at i,j, G(i,j ) is the green brightness value of the original true color image at i, j, and B(i, j) is the blue brightness value of the original true color image at i, j;

B, image filtering processing: adopt median filtering algorithm to carry out median filtering processing to the grayscale image that obtains through step A processing, obtain the grayscale image that removes noise information;

C, image binarization processing: adopt binarization threshold value segmentation method and carry out image segmentation processing to the gray scale image that removes noise information that step B obtains by f(x) function, the each pixel point in the gray scale image The grayscale value is compared with the optimal threshold T one by one, if the grayscale value of the pixel is less than the optimal threshold T, the grayscale value of the pixel is set to black, and the pixel with each grayscale value is black The area formed by the dots is marked as the background; if the gray value of the pixel is not less than the optimal threshold T, the gray value of the pixel is set to white, and the gray value of each pixel is white to form The white connected blocks marked as the image of the pill capsule; among them, T is the optimal threshold; the image binarization process in step C is to corrode the image of the image adhesion between every two tablet capsules in the grayscale image, and the image adhesion that is corroded is blended with the black background Integrate, so that the images of the two-adhesive tablet capsules are separated, and the images of each tablet capsule are separated to form a white connected block, which ensures the subsequent image edge detection and contour feature extraction of the image after step C processing. Image processing efficiency and counting accuracy for statistical counting;

Step 3: Image edge detection and contour feature extraction: use the Rober edge detection algorithm to perform image edge detection on the image of each tablet capsule in the grayscale image obtained in step C and extract the edge of the image of each tablet capsule, and obtain the image containing each tablet capsule. A grayscale image of the contour features of an image of a tablet capsule;

Step 4: Statistical counting: Use MATLAB tools to identify white connected blocks on a black background and count the number of areas where white connected blocks are output, so as to count the number of white connected blocks and obtain the number of tablets and capsules.

Due to the adoption of the above technical scheme, the present invention has remarkable technical effects: it adopts a CCD camera mechanism that can adjust the height position up and down and the horizontal position can be adjusted back and forth, and the inclination angle adjustment function of the bulk material plate through the inclination adjustment mechanism. The vibrating action of the vibration mechanism on the bulk material plate makes several tablets and capsules arranged and accommodated in the bottom of the W-shaped groove to form a row of tablets and capsules, which greatly improves the accuracy of collecting and obtaining true-color images of tablets and capsules. quality; using MATLAB tools, image processing is performed on the true-color images of the tablets and capsules by means of image processing, and accurate number counting of the tablets and capsules is carried out according to the image processing results, which greatly improves the accuracy and work efficiency of the counting of the tablets and capsules; and the device The overall design is reasonable, the structure is compact, the limited plane and space are fully utilized, the cost is low, and it is suitable for popularization and application.

Description of drawings

FIG. 1 is a schematic structural view of an embodiment of a machine vision-based tablet and capsule counting device of the present invention.

Fig. 2 is a partial structural schematic diagram of an embodiment of the present invention in which several tablet capsules to be counted are accommodated in the bottom of the W-shaped groove of the bulk material plate and form a row of regularly arranged tablet capsule queues.

Fig. 3 is a structural diagram of an embodiment of the assembly of the vibrating mechanism, the bulk plate, and the inclination adjusting mechanism of the present invention.

Fig. 4 is the original true-color image captured by the present invention through an industrial CCD camera that is housed in the groove bottom of the W-shaped groove and forms a row of regularly arranged rows of tablet capsules.

FIG. 5 is a grayscale image obtained after performing image grayscale processing on the original true-color image in FIG. 4 through Step A of Step 2.

FIG. 6 is a grayscale image obtained after performing image filtering on the grayscale image in FIG. 5 through Step B of Step 2.

FIG. 7 is a black-and-white image obtained after image binarization processing is performed on the grayscale image in FIG. 6 through Step C of Step 2.

FIG. 8 is a black and white image obtained after performing image edge detection and contour feature extraction on the black and white image in FIG. 7 through step three.

Detailed ways

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

A tablet capsule counting device based on machine vision, as shown in Figures 1-3, includes a chassis 15, a bulk material mechanism fixed in the chassis 15 and used to accommodate tablet capsules 16, and a CCD camera mechanism fixed on the top of the chassis 15 1. The upper computer module 1 connected with the CCD camera mechanism, the display screen 2 connected with the upper computer module 1, the top of the casing 15 is provided with a window 150 facing the CCD camera mechanism above, and the bulk material mechanism includes a tilt adjustment mechanism. The bulk material plate 11 fixedly arranged under the window 150 is tilted. The bulk material plate 11 is provided with several W-shaped grooves 110, and several tablet capsules 16 are arranged and accommodated in the groove bottom of the W-shaped groove 110 and form a row. Tablets and capsules 16 are regularly arranged in rows; the CCD camera mechanism includes an industrial CCD camera 9 that can adjust the height position up and down or adjust the horizontal position forward and backward. The industrial CCD camera 9 is fixed on the top of the chassis 15 and faces the window 150 below; Through the industrial CCD camera 9, a row of regularly arranged tablet capsules 16 queues formed in the groove bottom of the W-shaped groove 110 are photographed and the original true-color image is obtained, and the industrial CCD camera 9 collects the original image The true color image is transmitted to the upper computer module 1, and the upper computer module 1 uses MATLAB tools to perform image processing on the original true color image through an image processing method, and according to the image processing result, several tablet capsules 16 in the original true color image are processed. Statistical counting obtains the number of tablet capsules 16 in the groove bottom of the W-shaped groove 110 , and the number of tablet capsules 16 is displayed on the display screen 2 .

In this embodiment, the angle adjustment range of the inclination adjustment mechanism for the inclination angle between the bulk material plate 11 and the horizontal plane is 15-30 degrees.

In this embodiment, the inclination adjustment mechanism includes a push rod 19, a screw 13 that is rotatably fixed at the bottom of the push rod 19 and drives the push rod 19 to move up and down, and the top of the push rod 19 is fixed on the rear side of the bulk material plate 11 bottom of.

In this embodiment, it also includes a bar-shaped light source 12 fixedly arranged on both sides of the bulk material plate 11, and the two oppositely arranged bar-shaped light sources 12 simultaneously emit light to the bulk material plate 11 and illuminate the W-shaped groove 110. Tablet capsules 16 queues that are regularly arranged in a row form in the groove bottom.

In this embodiment, it also includes a vibrating mechanism for vibrating the bulk material plate 11. The vibrating mechanism includes a vibrating shaft 17, a motor 18 that drives the vibrating shaft 17 to vibrate, and the top of the vibrating shaft 17 is fixed on the bottom of the bulk material plate 11. In the center, the motor 18 is fixedly located at the bottom of the vibrating shaft 17 .

In this embodiment, the CCD camera mechanism also includes a support rod 3 fixedly arranged on the rear side of the cabinet 15, a connecting block 4 fixedly arranged on the support rod 3 which can adjust the height position up and down or the horizontal position can be adjusted forward and backward, and fixedly connected to the connecting block 4. On the block 4 and the camera clip 7 that stretches out towards the window 150 direction, the up and down fine-tuning knob 5 and the front and rear fine-tuning knob 6 that are rotatably fixed between the connecting block 4 and the support rod 3 through the threaded hole, the industrial CCD camera 9 passes through The fastening knob 8 is fixedly arranged in the camera clip 7; the fine-tuning knob 5 is rotated up and down through the threaded hole, so that the connecting block 4 moves up and down along the support rod 3 and drives the camera clip 7 and the industrial CCD camera 9 to move up and down; Move the fine-tuning knob 6 back and forth, so that the connecting block 4 moves back and forth along the support rod 3 and drives the camera clip 7 and the industrial CCD camera 9 to move back and forth.

In this embodiment, one baffle plate 10 is fixed on the front and rear sides of the bulk material plate 11 .

In this embodiment, the inclination adjustment mechanism also includes a rotatable hand handle 14 that is fixed on the lead screw 13 and drives the lead screw 13 to rotate. Of course, the hand handle can also be changed into an electric handle to realize automatic shaking. .

A method for counting tablets and capsules based on machine vision, comprising the steps of:

Step Ⅰ: Rotate the lead screw 13 through the hand handle 14 first, and the lead screw 13 drives the ejector rod 19 to move up or down at the same time, and the ejector rod 19 moving up or down simultaneously drives the rear side of the bulk material plate 11 to rise or fall. Thus the bulk material plate 11 is in the state of low front and high rear and the inclination angle between the bulk material plate 11 and the horizontal plane is 15-30 degrees; then several tablet capsules 16 to be counted are put into the bulk material plate 11, and then start The motor 18, the motor 18 drives the vibrating shaft 17 to vibrate the bulk material plate 11, and several tablet capsules 16 slide into the W-shaped grooves 110 under the action of the vibration until a row is formed in the bottom of each W-shaped groove 110 16 rows of pill capsules in row shape;

Step II: Then turn on two opposite bar-shaped light sources 12 at the same time, and the two oppositely arranged bar-shaped light sources 12 simultaneously emit light to the bulk material plate 11 and illuminate a row of queues formed in the bottom of the W-shaped groove 110 16 rows of pill capsules in row shape;

Step Ⅲ: Restart the industrial CCD camera 9, take pictures of the rows of pill capsules 16 queues illuminated by the industrial CCD camera 9 and obtain the original true-color image, as shown in Figure 4, the industrial CCD camera 9 again The collected original true-color image is transmitted to the upper computer module 1, and the upper computer module 1 uses MATLAB tools to perform image processing on the original true-color image through an image processing method, and performs image processing on the original true-color image according to the image processing result. Several tablet capsules 16 are statistically counted to obtain the number of tablet capsules 16 in the groove bottom of the W-shaped groove 110 , and the number of tablet capsules 16 is displayed on the display screen 2 .

In this embodiment, the image processing method in step III includes the following steps:

Step 1: Use MATLAB tools to read the original true-color image;

Step 2: Image preprocessing: including the following specific steps:

A. Image grayscale processing: According to the components of R, G, and B of each pixel in the original true-color image, the weight of each pixel is calculated through the Gray(i,j) function with R, G, and B as weights. Grayscale, so as to convert the original true color image into a grayscale image, as shown in Figure 5, where Gray(i,j)=0.299×R(i,j)+0.587×G(i,j)+ 0.114×B(i,j), Gray(i,j) is the gray level of the grayscale image at i,j, R(i,j) is the red brightness value of the original true color image at i,j , G(i, j) is the green brightness value of the original true color image at i, j, B(i, j) is the blue brightness value of the original true color image at i, j;

B, image filter processing: adopt median filter algorithm to carry out median filter processing to the grayscale image that obtains through step A processing, obtain the grayscale image that removes noise information, as shown in Figure 6;

C, image binarization processing: adopt binarization threshold value segmentation method and carry out image segmentation processing to the gray scale image that removes noise information that step B obtains by f(x) function, the each pixel point in the gray scale image The grayscale value is compared with the optimal threshold T one by one, if the grayscale value of the pixel is less than the optimal threshold T, then the grayscale value of the pixel is set to black, and the pixel with each grayscale value is black The area formed by the dots is marked as the background; if the gray value of the pixel is not less than the optimal threshold T, the gray value of the pixel is set to white, and the gray value of each pixel is white to form The white connected blocks marked as the image of the tablet capsule 16; where, T is optimal threshold; Obtain the image as shown in Figure 7 after the image binarization processing of step C;

Step 3: image edge detection and contour feature extraction: use the Rober edge detection algorithm to perform image edge detection on the image of each tablet capsule 16 in the grayscale image obtained in step C and extract the edge of the image of each tablet capsule 16 to obtain A black and white image of the outline features of the image containing each tablet capsule 16, as shown in Figure 8;

Step 4: Statistical counting: use the MATLAB tool to identify the white Unicom blocks under the black background and count the number of areas where the white Unicom blocks are output, so as to obtain the number of white Unicom blocks and obtain the number of tablets and capsules 16.

In a word, the above descriptions are only preferred embodiments of the present invention, and all equivalent changes and modifications made according to the scope of the patent application of the present invention shall fall within the scope of the patent of the present invention.

Claims (10)

1. A tablet capsule counting device based on machine vision, characterized in that: comprise a cabinet (15), be fixedly arranged in the cabinet (15) and be used to accommodate the bulk material mechanism of the tablet capsule (16), be fixedly arranged in the cabinet (15) ) above the CCD camera mechanism, the host computer module (1) connected with the CCD camera mechanism, the display screen (2) connected with the host computer module (1), and the top of the chassis (15) is equipped with a CCD camera mechanism facing the top The window (150), the bulk material mechanism includes The bulk material plate (11) arranged below the window (150) can be tiltably fixed by the inclination adjustment mechanism, The bulk material plate (11) is provided with several W-shaped grooves (110), and several tablet capsules (16) are arranged and accommodated in the bottom of the W-shaped grooves (110) to form a row of regularly arranged tablets. The capsule (16) queue; the CCD camera mechanism includes an industrial CCD camera (9) that can adjust the height position up and down or can adjust the horizontal position forward and backward. ) is facing; take pictures of a row of tablet capsules (16) formed in the groove bottom of the W-shaped groove (110) by an industrial CCD camera (9) and obtain an original true-color image. CCD camera (9) transmits the original true-color image that gathers to host computer module (1) again, and upper computer module (1) adopts MATLAB instrument and carries out image processing to original true-color image by image processing method, according to image The processing result counts several tablet capsules (16) in the original true-color image and obtains the number of tablet capsules (16) in the groove bottom of the W-shaped groove (110), and the number of tablet capsules (16) is shown on the display (2). 2. A kind of tablet capsule counting device based on machine vision according to claim 1, characterized in that: the angle adjustment range of the inclination adjustment mechanism to the inclination angle between the bulk material plate (11) and the horizontal plane is 15- 30 degrees. 3. A kind of tablet capsule counting device based on machine vision according to claim 1 or 2, characterized in that: the inclination adjustment mechanism comprises a push rod (19), which is rotatably fixed at the bottom end of the push rod (19). And drive the leading screw (13) that push rod (19) moves up and down, the top of push rod (19) is fixedly located at the bottom of bulk plate (11) rear side. 4. A kind of tablet capsule counting device based on machine vision according to claim 3, characterized in that: it also includes a strip light source (12) fixedly arranged on both sides of the bulk material plate (11), two opposite The set strip light source (12) simultaneously emits light to the bulk material plate (11) and illuminates a row of regularly arranged tablet capsules (16) formed in the groove bottom of the W-shaped groove (110). 5. according to claim 1 or 2 or 4 described a kind of tablet capsule counting device based on machine vision, it is characterized in that: also comprise the vibrating mechanism that vibrates bulk material plate (11), and vibrating mechanism comprises vibrating shaft ( 17), the motor (18) that drives the vibration shaft (17) to generate vibration, the top of the vibration shaft (17) is fixed at the center of the bottom of the bulk material plate (11), and the motor (18) is fixed at the bottom of the vibration shaft (17) end. 6. A kind of tablet capsule counting device based on machine vision according to claim 5, characterized in that: the CCD camera mechanism also includes a support rod (3) fixedly arranged on the rear side of the cabinet (15), the height position of which can be adjusted up and down Or the connection block (4) that can be fixedly arranged on the support rod (3) that can adjust the horizontal position back and forth, the camera clip (7) that is fixedly connected on the connection block (4) and stretches out to the window (150) direction, through the screw thread The hole can be rotated to fix the upper and lower fine-tuning knobs (5) and the front and rear fine-tuning knobs (6) between the connecting block (4) and the support rod (3). The industrial CCD camera (9) is fixed by the fastening knob (8) Set in the camera clip (7); turn the up and down fine-tuning knob (5) through the threaded hole, so that the connecting block (4) moves up and down along the support rod (3) and simultaneously drives the camera clip (7) and the industrial CCD camera (9) Move up and down; turn the front and rear fine-tuning knobs (6) through the threaded holes, so that the connecting block (4) moves back and forth along the support rod (3) and simultaneously drives the camera clip (7) and the industrial CCD camera (9) to move back and forth. 7. A tablet and capsule counting device based on machine vision according to claim 1 or 2 or 4 or 6, characterized in that: the front and rear sides of the bulk material plate (11) are fixed with a baffle plate (10). 8. A machine vision-based tablet and capsule counting device according to claim 7, characterized in that: the inclination adjustment mechanism also includes a rotatable device fixed on the lead screw (13) and driving the lead screw (13) to rotate Hand handle (14) or electric handle. 9. A tablet capsule counting method based on machine vision, characterized in that: comprises the steps: Step Ⅰ: Rotate the lead screw (13) through the hand handle (14), and the lead screw (13) drives the ejector rod (19) to move up or down at the same time, and the ejector rod (19) moving up or down simultaneously drives The rear side of the bulk material plate (11) rises or falls, so that the bulk material plate (11) is in a state of low front and high rear and the angle of inclination between the bulk material plate (11) and the horizontal plane is 15-30 degrees; Several tablet capsules (16) are put into the bulk material plate (11), then start the motor (18), and the motor (18) drives the vibrating shaft (17) to produce a vibration effect on the bulk material plate (11), and several tablet capsules (16) slide into the W-shaped groove (110) under the action of vibration until forming a row of rows of tablet capsules (16) in the groove bottom of each W-shaped groove (110); Step II: Then turn on two oppositely arranged bar-shaped light sources (12) at the same time, and the two oppositely arranged bar-shaped light sources (12) simultaneously emit light to the bulk material plate (11) and illuminate the W-shaped groove (110) Tablet capsules (16) queues formed in rows in the groove bottom; Step Ⅲ: Restart the industrial CCD camera (9), take a picture of the tablet capsules (16) queues that are illuminated in rows and rows by the industrial CCD camera (9) and obtain the original true color image, the industrial CCD camera (9) ) and transmit the collected original true-color image to the upper computer module (1), the upper computer module (1) adopts MATLAB tool and carries out image processing to the original true-color image by image processing method, according to the image processing result to the original Several tablet capsules (16) in the true-color image of the statistic count obtain the number of tablet capsules (16) in the groove bottom of W-shaped groove (110), and the number of tablet capsules (16) is shown on the display screen (2) up. 10. A kind of tablet capsule counting method based on machine vision according to claim 9, is characterized in that: the image processing method in step III comprises the steps: Step 1: Use MATLAB tools to read the original true-color image; Step 2: Image preprocessing: including the following specific steps: A. Image grayscale processing: According to the components of R, G, and B of each pixel in the original true-color image, the weight of each pixel is calculated through the Gray(i,j) function with R, G, and B as weights. Grayscale, so as to convert the original true color image into a grayscale image, where Gray(i,j)=0.299×R(i,j)+0.587×G(i,j)+0.114×B(i, j), Gray(i,j) is the gray level of the grayscale image at (i,j), R(i,j) is the red brightness value of the original true color image at (i,j), G (i, j) is the green brightness value of the original true color image at (i, j), and B(i, j) is the blue brightness value of the original true color image at (i, j); B, image filtering processing: adopt median filtering algorithm to carry out median filtering processing to the grayscale image that obtains through step A processing, obtain the grayscale image that removes noise information; C, image binarization processing: adopt binarization threshold value segmentation method and carry out image segmentation processing to the gray scale image that removes noise information that step B obtains by f(x) function, the each pixel point in the gray scale image The grayscale value is compared with the optimal threshold T one by one, if the grayscale value of the pixel is less than the optimal threshold T, the grayscale value of the pixel is set to black, and the pixel with each grayscale value is black The area formed by the dots is marked as the background; if the gray value of the pixel is not less than the optimal threshold T, the gray value of the pixel is set to white, and the gray value of each pixel is white to form The white connected block mark is the image of the tablet capsule (16); wherein, T is the optimal threshold; Step 3: image edge detection and contour feature extraction: use the Rober edge detection algorithm to perform image edge detection on the image of each tablet capsule (16) in the grayscale image obtained in step C and extract the image of each tablet capsule (16) The edge of obtains the black-and-white image that comprises the contour feature of the image of each tablet capsule (16); Step 4: Statistical counting: Discriminate the white Unicom blocks under the black background by MATLAB tool and count the area quantity of the output white Unicom blocks, thereby counting the number of white Unicom blocks, thereby obtaining the number of tablets and capsules (16) number.