KR20070071733A - A quality monitoring system of butchered chicken line and method thereof - Google Patents

Abstract

A system and a method for controlling quality in a chicken butchery line are provided to utilize check result data/image to each process management by automatically checking diseases/wounds received in a farming process and the wounds received in each process through image recognition, treatment, and process using a camera installed to each line. Each camera(203a,203b) captures a picture of the butchered chicken(100) moved along a rail. A checker(400) extracts an image from the captured picture, detects a quality level and defect causes through weight measurement and shape/color analysis by using volume estimation based on image processing, and integrates results. A data storing server(500) stores and manages the integrated analysis results of the checker by using a database. An image storing server(600) manages the image received from the checker by using the database and includes a function as a terminal client server. A plurality of clients(700A-700N) control/monitor a process at each process of the line, and provide image data analyzed in the checker by connecting to the image storing server.

Description

Quality inspection system and method in the manufactory process {A QUALITY MONITORING SYSTEM OF BUTCHERED CHICKEN LINE AND METHOD THEREOF}

1 is a block diagram of a quality inspection system in a yard process according to the present invention.

2 is a detailed configuration diagram of the inspection apparatus in FIG.

3 is a flowchart of an embodiment of performing a quality inspection in a metric process according to the present invention.

4 is a flow chart for image acquisition in FIG.

5 is a flowchart illustrating the generation of a weight learning table for quality inspection in a metric process according to the present invention.

6 is a flowchart illustrating the creation of a reference learning table for quality inspection in a metric process according to the present invention.

7 is a flow chart for a typical pottery production process.

<Description of Symbols for Main Parts of Drawings>

100: total 101: hanger

201: sensor 203: camera

301: high frequency ballast 303: first light source

305: second light source 400: inspection device

500: data storage server 600: image storage server

700A-700N: Client Device

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pottery process, and more particularly, to objectively and uniformly inspect a product by automatically checking the quality of livelihood through image recognition in the pottery process. It relates to a metrology quality inspection system and method for obtaining a phosphorus evaluation.

In general, a process for producing a pottery system will be described with reference to FIG. 7.

A poultry farm (B) operated directly by the company producing the poultry or strategically contracted with the poultry producer is loaded onto a vehicle (A) dispatched according to the release plan of livestock raised. And (c) deliver it to the poultry producer (C).

The poultry producers are weighed in order to pay the poultry farms (B) for the livelihood received (D) and then enter the poultry process (E).

In the flotation process, the livelihood of the input is stunned by a predetermined method, and then a blood bleeding process is performed in which blood in the livelihood is discharged (F).

Subsequently, in order to extract the hair of the bleeding livelihood, the process of separating the hair with a predetermined device using the principle of centrifugal separation and then (G).

As described above, the bleeding and the hair-separated sludge are transferred to the next process through the transfer rail (H) and the declining process is performed (I) The extraction process of separating by-products such as internal organs is performed. (J).

When the extraction process is completed, the foreign matters on the shedding system are completely removed through the washing process, and then the cooling process is carried out to maintain the freshness (L) and transferred to the packaging process. It is supplied to the consumer (M).

In producing the yardmeter through the above process, the inspection for the uniformity of quality is carried out, in the case of the inspection is dependent on the eyes of skilled workers.

However, the metrology line is an automated system operating at a speed of 2.5 birds per second, and there is a limit to the entire inspection of the quality of the metrology through the naked eyes of the operator, so that only the sample inspection is completed after the completion of the process. Due to this, there is a disadvantage in that the homogeneity of the inspection and the fairness of the quality grade cannot be secured.

In addition, due to the quality inspection based on the subjective judgment of each worker has a disadvantage that there are a lot of difficulties in tracking the cause of poor quality and poor quality of livelihood supplied to the consumer.

In addition, since the sample inspection by the operator's visual inspection is not a whole inspection, there is a possibility that the defective product is made into a final product, which causes problems in quality assurance, and the consistency of the inspection specification is lost depending on the skill of the operator and the condition on the day. Disadvantages arise.

The present invention has been invented to solve the above problems, the object of which is to install a camera on the line in the slaughterhouse factory, through the image recognition, processing and processing of livelihood disease or wound or car or transportation process The system automatically inspects wounds and wounds in the process and stores test result data and images for use in each process control.

In other words, through the analysis of the automatic inspection results, it is possible to estimate the causes of defects in livestock raising, loading, transporting, and slaughtering processes, and ultimately to provide systematic breeding management such as improvement of breeding quality and prevention of deterioration during loading or transportation. will be.

In addition, through a full inspection in the process of phasing, the quality assurance system is established by preventing the distribution of defective products due to sample inspection, providing accurate data on product grade determination, the parts generated when equipments occur, the reason for the grade determination, and the corresponding products. To provide an image.

In addition, it performs volume detection, search for equipment through color, and search for equipment through form on average of three products per second. It makes it possible to make an accurate estimation and to easily analyze the status of equipment occurrences by livestock farms and transportation personnel.

Then, it is to provide uniformity and objectivity of inspection by inspecting the quality of the meter through image processing.

The present invention for realizing the above object is an image acquisition means for acquiring an image of the yard disposed at a predetermined position of the yard process line and moved through the rail; Extracting the image from the image of the metric applied by the image acquisition means, through the processing and processing to detect the quality grade and the cause of the defect by weighing through the analogy of the volume, analysis of the shape and color, and synthesize the results An inspection apparatus; A data storage server configured to store and manage data of a comprehensive analysis provided by the inspection apparatus; An image storage server managing image data provided by the inspection apparatus as a database, the image storage server including a function of a terminal client server; A process control and monitoring apparatus disposed in each process of a metric line, comprising: a plurality of client apparatuses connected to the image storage server via Ethernet to provide monitoring of image data analyzed by the inspection apparatus. Provides a metric quality inspection system.

The present invention also provides a method for acquiring a stereoscopic image of a metric system in an image acquisition area of a metric line; Extracting only an image of a metric, which is a target object, by separating the background from the acquired image; Extracting the extracted image by dividing the extracted image into parts; Estimating the weight of the meter through a set learning table by measuring the total amount of pixels of the extracted image; Analyzing the R, G, and B colors of the extracted image to detect quality inspection and shape defects; And a process of generating the extracted weight data, the quality inspection and the shape defect result, and storing the result data and the image data.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

As shown in FIG. 1, the metrology quality inspection system according to the present invention includes a panel 90, a hanger 101, a sensor 201, a camera 203, a high frequency stabilizer 301, and a first light source 303. And an image acquiring means including a second light source 305, an inspection apparatus 400, a data storage server 500, an image storage server 600, and a client apparatus 700.

The image acquiring means is disposed on a line for transferring the tapping completed in the automated pottery line to the cutting process.

The hanger 101 in the image acquiring means is moved through a rail in a system line composed of an automated system, and after the wet, the group of the hairs 100 separated from the hair is hung and transferred to the next process, the cutting process.

The sensor 201 is mounted at a predetermined position on the rail of the duct line and senses a transfer position of the hanger 101 on which the duct is caught and outputs information on the transfer position.

The camera 203 is a high-performance digital camera, which is disposed on the front side and rear side of the line, respectively, and is a pedometer that is transferred through the hanger 101 in synchronization with a trigger signal applied from the sensor 201. Acquire images in three dimensions.

The first light source 303 and the second light source 305 are arranged as a high frequency fluorescent lamp, and according to the stable voltage applied from the high frequency ballast 301 to ensure the same level of illumination at all times without flickering to the image acquisition area of the pedometer Give it.

The panel 90 is disposed with a black background so that only the metric, which is a target object, can be easily extracted from the image acquired by the camera 203.

Inspection device 400 is disposed on the front side and the rear side of the line to extract the image from the image acquired from the camera 203 through the processing and processing, weighing through volume inference, analysis of the shape and color through It classifies the quality grade and the cause of the defect and aggregates the analyzed results to be used as feedback data.

The inspection device 400 registers the defect detection reference data for each breed / line of the yard, accesses the defect detection reference data according to the input schedule, and performs a defect inspection matching the livelihood.

The data storage server 500 manages the data obtained from the comprehensive analysis provided by the inspection apparatus 400.

The image storage server 600 manages the image data provided by the inspection apparatus 400 as a database, and includes each of the client devices 700A to 700N involved in the process of the metering process through the Ethernet, including the function of the terminal client server. Connect to enable monitoring of the image data analyzed by the inspection apparatus 400.

Client device 700A to 700N A process control device and a monitoring device arranged in each process of the metering line are connected to the image storage server 600 through Ethernet to monitor the results of the quality inspection analyzed by the inspection device 400. do.

As shown in FIG. 2, the inspection apparatus 400 includes an interface unit 401, an image extraction unit 403, a site extraction unit 405, a total pixel measurement unit 407, a color analyzer 409, It is configured to include a fracture detection unit 411, a control unit 413, a memory unit 415 and the communication unit 417.

The interface unit 401 interfaces the phantom image in frame units acquired by the cameras 203 disposed on the front and rear sides of the line, respectively.

The image extractor 403 extracts only the image of the metric from the image applied through the interface unit 401.

In extracting the image, a threshold value is applied to process a pixel above a threshold value as '1', and a pixel below the threshold value is processed as '0' to extract an image of a metric system as a target object, and the extracted image. Since the noise component may be included in the image, a painting algorithm is applied to calculate the number of constituent pixels.

In other words, the area consisting of the largest chunk is recognized as a target object by painting the connected area while adding the number of pixels to extract the image of only the noise-free material.

The site extraction unit 405 classifies a site for weight measurement and quality analysis in the extracted metric image.

Extraction of the specific area divides the image of the physiognostic object, which is a target object, into 10 parts such as the left leg, the right leg, the abdominal cavity, the left abdomen, the abdomen, the right abdomen, the left wing, the chest, the right wing, and the neck.

Then, the horizontal deviation is obtained along the outline of the object while the vertical deviation is obtained, and the part where the deviation rapidly changes is used as a reference for dividing the parts.

For example, if you find the crotch, if you move in the horizontal direction while differentiating (from left to right), if the vertical deviation is lowered and becomes higher, then the peak becomes the point where the crotch is divided.

The total pixel measuring unit 407 may measure the total pixel of the segmented image to estimate the weight of the corresponding meter.

That is, a certain amount of weight is applied to the sum of the pixels excluding the wing and the neck, which are relatively large deviations.

The color analyzer 409 may estimate the quality and shape of the corresponding metrology by analyzing R, G, and B colors in the image from which only the specific region is extracted.

Fracture detection unit 411 generates a detection line next to the flank of the extracted image and if the wing is not detected in the detection line can be determined as a shape defect due to fracture or dislocation.

The control unit 413 estimates the weight of the gauge from the weight table set through learning about the total pixels applied by the total pixel measuring unit 407, and the colors of R, G, and B applied by the color analyzer 409. The quality is determined from the signal, and the shape of the gage is determined from the signal applied from the fracture detector 411.

That is, when all pixels of the part recognized as a metric are traversed, when the pixel meets the color, size, and area that are registered detection criteria, the mask is masked. Poor grades (A, B, C, D grades) such as chickens, carnival 1, carnival 2, red chickens, skin breakage, etc., and areas of occurrence (left leg, right leg, abdominal cavity, left abdomen, abdomen, right abdomen, left wing, chest) , Right wing, neck), color of defect (R, G, B), size of defect (mm), etc. are determined.

Then, a fracture detection reference line is set below a certain line of the metric system, and when a wing falls below the fracture detection line, it is recognized as a fracture.

In other words, when a fracture or dislocation occurs, the wing sags downward. Thus, a fracture detection line is generated next to the neck area, and when it is detected that the wing sags below the fracture detection line, it is determined as fracture or dislocation.

In addition, a fracture detection line is generated next to the flank to detect a case in which the wing is completely broken and comes into contact with the neck or is torn off.

The memory unit 415 stores a weight learning table and an inspection reference table set through data and learning necessary for operating the system, and temporarily stores weight estimation and quality inspection results.

The communication unit 417 is configured as, for example, a LAN card, and is connected to the data storage server 500 and the image storage server 600 to transmit weight estimation and quality inspection result data analyzed by the inspection apparatus 400. .

Referring to the operation of automatically checking the quality of the meter by applying the present invention including the function described above as follows.

The first light source 303 and the second light source 305 ensure optimal lighting in the image acquisition line according to the voltage applied from the high frequency ballast 301.

When the inspection is started while the illumination for the image acquisition is secured as described above, the defective reference data registered for each breed (chicken or chicken) / line of the breeder is accessed according to the input schedule to match the meter.

Subsequently, two cameras 203 arranged in two of the image acquisition lines of the metrology lines acquire a stereoscopic image of the pedometer 100 which is suspended from the hanger 101 (S100).

That is, as can be seen in Figure 4, in the process of moving through the rail hanging on the hanger 101, the detached pedometer in the bathing process (S101) when the movement position is detected by the sensor 201 (S102), the sensor 201 outputs a trigger signal for image acquisition to the high performance digital camera 203 disposed at the front side and the rear side of the image acquisition line (S103).

Accordingly, the two cameras disposed at the front side and the rear side are operated in synchronization with the trigger signal applied from the sensor 201 (S104), thereby acquiring a stereoscopic image of the pedometer 100 suspended in the rinsing 101 (S105). ).

As described above, when the two cameras 203 acquire the stereoscopic image of the metric system 100, the two cameras 203 transmit the same to the inspection apparatus 400 in real time.

The inspection apparatus 400 extracts a pure image of only an object, that is, a pedometer, from a stereoscopic image applied from two cameras 203 (S200).

In the extraction of the image, a threshold value is applied to process a pixel above a threshold in a stereoscopic color image of an input colorimeter as '1', and a pixel below a threshold value is processed as '0' to extract an image of the metrology. Since a noise component may be included in the extracted image, a painting algorithm is applied to calculate the number of constituent pixels.

In other words, by painting the connected area, the number of pixels is added to extract the area consisting of the largest chunk as the image of the spectral system to remove the ambient noise.

When the image of the meter is extracted as described above, the parts for weight measurement and quality analysis are distinguished (S300).

The division of the part is divided into a total of 10 sites such as left leg, right leg, abdominal cavity, left abdomen, abdomen, right abdomen, left wing, chest, right wing, neck.

Then, the horizontal deviation is obtained along the outline of the object while the vertical deviation is obtained, and the part where the deviation rapidly changes is used as a reference for dividing the parts.

For example, if you find the crotch, if you move in the horizontal direction while differentiating (from left to right), if the vertical deviation is lowered and becomes higher, then the peak becomes the point where the crotch is divided.

In this way, the total pixels are measured to calculate the weight of the metric system for the divided images.The total pixels are applied by applying a certain amount of weight to the sum of the pixels excluding the wing and the neck, which are relatively large deviations. The measurement estimates the overall volume, not just the area of the meter (S400).

When the estimation of the total volume is completed by measuring the total pixels, the weight of the corresponding meter is estimated by applying the weight learning table set through learning (S500).

The weight estimation of the metric is to obtain the average weight and volume data for the breeding farm, and the set weight learning table is set through an iterative operation of matching the actual weight and the total amount of pixels measured for the real as shown in FIG. do.

The weight learning table is set as shown in Table 1 below, for example.

⊙ Classification division Broiler division Three worlds Weight (g) Weight (g) No. 5 501-600 No. 30 301-350 No. 6 601-700 No. 35 351-400 No. 7 701-800 No. 40 401-450 No. 8 801 to 900 No. 45 451 to 500 No. 9 901 to 1000 No. 50 501-550 No. 10 1001-1100 No. 55 551 to 600 No. 11 1101-1200 No. 60 601-650 No. 12 1201-1300 Bag 078 651-800 No. 13 1301-1400 Bag 089 801 to 900 No. 14 1401-1500 Bag091 901 to 1000 No. 15 1501-1600 Bag101 1000-1500 No. 16 1601-1700 Out of standard 300 or less No. 17 1701-1800 18-25 1801 or higher Out of standard 500 or less

⊙ Body standard

Classification Classification standard C grade B grade A grade Broiler 500g or less 500-600 g More than 600g Three worlds 300 or less 300-350 g 350g or more

When the weight is calculated from the total pixel amount according to the above criteria, the R, G, and B colors of the extracted image are analyzed (S600) to determine whether the quality and shape of the corresponding meter are defective (S700).

That is, the purpose of the present invention is to improve quality by detecting wounds and diseases occurring in breeding, loading, transporting, and slaughtering processes by analyzing R, G, and B colors on extracted culling images.

The determination of the R, G, and B colors is performed by comparing the red, green, and blue values of the pixels with the detection reference data for the color and summing the number of pixels meeting the detection criteria. If an abnormality is found for the defect, information about the grade, size, R / G / B value, and the area with the largest mass is generated.

In S700, a reference table is set to determine whether the quality and shape of the meter are defective. The reference table is generated and set through a color reference measured for a real object and an iterative simulation applied thereto as shown in FIG. 6.

The determination of the quality and shape of the metric system through the analysis of the R, G, and B colors is performed by circulating all the pixels of the part recognized as the metric, masking when the pixel meets the color, size, and region, which are registered detection criteria. Using the masked part, the grades of defects (A, B, C, D grades) such as bruises (discoloration), navel chickens, knee chickens, carnival 1, carnival 2, fowl, skin breakage, etc. Determine left leg, right leg, abdominal cavity, left abdomen, abdomen, right abdomen, left wing, chest, right wing, neck), color of defect (R, G, B), size of defect (mm).

The method of detecting a bruise in the name of the defect is as follows.

The bruise (discoloration) caused by external shock is mainly caused by farms or transportation process, because if you hang at the factory and bleeding, you can not bruise even if you hit another place.

Therefore, if there is a bruise on a specific part of the yard, the value of R or B of a certain part is higher than that of a normal yard. Check and use as data for breeding management and transportation management.

Detection of the bruise (discoloration) is extracted by applying the criteria as shown in Table 2, for example, depending on the size, number and degree of darkening.

Bruise Type of bruise standard C grade B grade A grade A dark bruise R / G / B 136/136/102 Size (mm) 20 * 20 19 or less * 10 or more 10 * 10 Count 1 or more 2 or more 1 or less Light bruise R / G / B 204/119/153 Size (mm) 30 * 30 29 or less * 29 or less 10 * 10 or less Count 3 or more 2 1 or less

In addition, the detection of navel chickens in the name of the defect applies the following method.

The navel chicken is a scab that is wounded by a cut in the chest during farming and usually appears black, dark red or dark yellow.

 Therefore, through the R, G, B information and size and number of the scab around the navel, for example, by applying the criteria as shown in Table 3 below.

Belly chicken Classification standard C grade B grade A grade R / G / B 221/204/184 221/204/184 Normal color Size (mm) 10 * 10 or more 10 * 10 or more radish Count 2 or more 1 or more radish

In addition, the detection of the knee chicken in the name of the defect applies the following method.

Knee chickens are a phenomenon that occurs when chickens sit for a long time due to poor farming environment, etc., and appear as black scabs around the knees.

Therefore, through the R, G, B information and size and number of the scab area around the knee, for example, it is extracted by applying the criteria as shown in Table 4 below.

Knee chicken Classification standard C grade B grade A grade R / G / B 170/119/170 Size (mm) 5 * 5 or more 5 * 5 or less radish Count 1 or more 1 or more radish

 In addition, the detection of Carnival 1 in the name of the defect applies the following method.

Carnival 1 is a scab caused by wounds that have been pecked on each other in a livelihood of a farm, mainly occurring around the anus and appear as dark or light scabs around the anus or the entire body.

Therefore, through R, G, B information and size and number of the scab area around the anus or torso, for example, by applying the criteria as shown in Table 5 below.

Ganibal 1 Classification standard C grade B grade A grade R / G / B 221/204/184 Size (mm) 5 * 5 or more 5 * 5 or less 5 * 5 or less Count 6 or more 5 to 2 2 or less

In addition, Carnival 2 in the name of the defective means a scab caused by a wound caused by a scratch on the hip or thigh side in the livelihood of the farmhouse, and appears as a dark or light scab.

Therefore, through the R, G, B information and the size and number of the scab area generated on the hip or thigh side, for example, by applying the criteria as shown in Table 6 below.

Ganibal 2 Classification standard C grade B grade A grade R / G / B 255/119/204 Size (mm) 2 * 30 or more 2 * 30 or more 2 * 30 or more Count 6 or more 2 to 4 1 or less

In addition, in the name of the defect, red gypsy is a phenomenon that appears that the bleeding is not good as it is put into the factory line, it appears closer to red than normal.

In other words, it must be alive and lived until it is put into the factory line, but it is a phenomenon that occurs when blood is not made normally when it is put into the line in an unhealthy or poor breathing state due to the farming state or transportation.

Therefore, the red ginseng is a red product as a whole, so, for example through the information of the R, G, B is extracted by applying the criteria as shown in Table 7 below, if it is determined as a red ginseng unconditionally classified as C grade .

Red chicken Classification standard C grade B grade A grade R / G / B 255/187/204 normal normal

In addition, the skin breakage in the name of the defect is a state in which the skin of the product is torn or broken, and appears red and long, and the skin breakage is different from the bruise (discoloration). Or the value of B is high.

Therefore, the skin damage is extracted by applying the criteria as shown in Table 8 below, for example, through the color gradient and the broken length when the R, G, B information of the damaged part is suddenly different from the surrounding pixels.

Skin breakage Classification standard C grade B grade A grade R / B 225/187 Color gradient Length 3 * 10 or more 3 * 10 or less radish

When the detection determination is made with respect to the shape defect of the metric system through the above process, the fracture detection and determination is performed (S800).

The fracture detection and determination applies a matching method with a pattern within an allowable value, for which a "Skeleton Morphological Pattern Matching" algorithm is used.

Dislocations and fractures are difficult to detect due to the comparison of simple image patterns due to the nature of the process. Therefore, the coupling force between pixels is compared and structural detection is performed.

 To this end, a fracture detection reference line is set below a certain line of the metric system, and when a wing falls below the fracture detection line, it is recognized as fracture.

In other words, when a fracture or dislocation occurs, the wing sags downward. Thus, a fracture detection line is generated next to the neck area, and when it is detected that the wing sags below the fracture detection line, it is determined as fracture or dislocation.

In addition, a fracture detection line is generated next to the flank to detect a case in which the wing is completely broken and comes into contact with the neck or is torn off.

Since the fracture and the dislocation itself is a defective product, if a fracture or dislocation is detected, it is classified as a C grade, the detection for this is applied by applying the criteria set as shown in Table 9, for example.

Dislocation / fracture Classification standard C grade A grade R / G / B 204/85/119 size Dislocation itself is bad Count 1 or more radish Matching degree

If the determination of the weight and quality of the slaughterhouse is made through image recognition through the above process, the cause of the quality abnormality is caused by the influence of the farming environment of the farm, the amount of feed, etc. Inspection result data and weight data are generated whether the result is due to the process or due to the line of the meter (S900).

The stored data is classified into the date and the failure occurrence status, the average volume distribution by farmhouse, the failure occurrence status by farmhouse, the failure occurrence status by transport person, etc., and outputted in report or graph form. To provide ease of analysis.

In addition, based on the passage of time, the status of occurrence of total defects, the occurrence of defects by farmers, and the occurrence of defects by transportation personnel are classified and stored.

Subsequently, the generated weight data and the test result data are stored in the data storage server 500 and the image storage server 600, so that they can be utilized as feedback data for livestock breeding and management in farms (S100).

Since the image storage server 600 and the client device disposed in each line of the metric process are connected via Ethernet, it is possible to monitor the defect occurrence status for each cause in each line.

Although the present invention has been described with reference to one embodiment shown in the drawings, this is merely exemplary, and it will be understood by those skilled in the art that various modifications and equivalent embodiments are possible.

As described above, the present invention provides an accurate analysis of the causes of deterioration and damage of livelihoods from breeding to livelihood by constructing and automating an image recognition system for the livelihood quality inspection in the inefficient and inaccurate shedding process by conventional visual inspection. It provides the effect of improving productivity.

By adopting a high-performance digital CCD camera to acquire, process, and process the image, it provides a weight measurement function through volume inference, quality grade through shape and color, and classification of defect cause, thereby improving the reliability of inspection.

In addition, by collecting, counting, inquiring and analyzing various livelihood quality information and data generated in the livestock process in real time, and linking livelihood grade data and image data to the breeding management system, dispatching and rooting management system, By laying the foundation for management, it reduces the manpower and enhances the reliability of inspection by automation, and improves productivity and provides the effect of securing the efficiency of breeding and transportation process through feedback of quality information.

Claims (25)

Image acquisition means for acquiring an image of a pedometer disposed at a predetermined position on the metric process line and moved through a rail; Extracting the image from the image of the metric applied by the image acquisition means, through the processing and processing to detect the quality grade and the cause of the defect by weighing through the analogy of the volume, analysis of the shape and color, and synthesize the results An inspection apparatus; A data storage server configured to store and manage data of a comprehensive analysis provided by the inspection apparatus; An image storage server managing image data provided by the inspection apparatus as a database, the image storage server including a function of a terminal client server; A process control and monitoring apparatus disposed in each process of a metric line, comprising: a plurality of client apparatuses connected to the image storage server via Ethernet to provide monitoring of image data analyzed by the inspection apparatus. Metrology quality inspection system. The method of claim 1, The image acquisition means is a quantitative quality inspection system, characterized in that disposed on the line for transferring the complete tangometer in the automated yard line to the cutting process. The method of claim 1, The image acquisition means includes a panel provided in the image acquisition area to provide a black background; A sensor disposed at a predetermined position on a rail of the automated meter line to sense a transfer position of the meter; A camera disposed in the image acquisition area for acquiring an image of a metric system which is transferred in synchronization with a trigger signal of a sensor; One or more light sources for providing illumination to the image acquisition area; And a voltage regulator providing a stable voltage to the light source. The method of claim 3, The camera is disposed on the front side and the rear side of the image acquisition area, respectively, the metric quality inspection system, characterized in that for acquiring the stereoscopic image of the metric system transferred in synchronization with the trigger signal of the sensor. The method of claim 3, And a plurality of cameras are provided in at least two or more in said image acquisition area to acquire stereoscopic images. The method of claim 1, The inspection apparatus is a quantitative quality inspection system characterized in that the defect detection reference data for each type / line of the metric system is registered, and access to the defect detection reference data according to the input schedule to perform a test matched to the metric system. The method of claim 1, The inspection apparatus provides the classification of each part from the pixels of the stereoscopic image of the acquired yard, and the color quality of the yard, which determines the defects generated in the farming process, the livelihood transportation, and the process of the shedding process. Inspection system. The method of claim 1, The inspection apparatus is a quantitative quality inspection system, characterized in that for determining whether the fracture or dislocation defects from the pixels of the stereoscopic image of the acquired metric. The method of claim 1, The inspection device masks when all pixels of the part recognized as a metric in the extracted image are met and meets the pixel matching the color, size, and area, which are registered detection criteria, and uses the masked part to name the defects of bruise and navel. Poor grades (A, B, C, D grades) for chickens, knee chickens, Carnival 1, Carnival 2, Red chickens, skin breakage and the site of defects (left leg, right leg, abdominal cavity, left abdomen, abdomen, right abdomen, Left wing, chest, right wing, neck), the color of the defect (R, G, B), the quality of the metrology system characterized in that it determines the size of the defect (mm). The method of claim 1, The image extraction in the inspection apparatus applies a threshold to process a '1' for pixels above a threshold and a '0' for a pixel below a threshold to extract an image of a pedometer as a target object. A metrology quality inspection system characterized by extracting an image of only a noise-free yard, by recognizing the area consisting of the largest mass by adding the number of pixels and painting the connected area as a yardmeter. The method of claim 1, The area extraction of the inspection device is obtained by differentiating horizontally and vertically along the outline of the image of the yard, based on dividing the site of the part where the deviation rapidly changes. , Quality measurement system, characterized in that divided into a total of 10 parts, right wing, left wing, chest, right wing, neck. The method of claim 1, The total pixel extraction of the inspection apparatus is extracted by applying a certain amount of weights to the sum of the pixels excluding the wing and the neck, which are the parts of which the deviation is large, thereby estimating the overall volume as well as the area of the yard. Quality inspection system. The method of claim 1, The color analysis of the inspection apparatus compares the R, G, and B values of the pixels with the detection reference data for the color, and adds the number of pixels that meet the detection criteria, and if a size error is found, the grade for the defective Metrology quality inspection system, characterized in that the information including the size, R / G / B value, the area with the largest mass and extracted as a test result. The method of claim 1, The fracture analysis of the inspection device is a quantitative quality inspection system, characterized in that the determination of the presence of the blade around the reference detection line. The method of claim 1, The inspection apparatus may include: an interface unit connected to an image acquisition unit to interface a phantom image in a frame unit; An image extracting unit extracting only an image from an image applied through the interface unit; A part extraction unit for classifying a part for weight measurement and quality analysis in the extracted metric image; A total pixel measuring unit configured to measure a total pixel of the segmented image to estimate the weight of the corresponding meter; A color analyzer for estimating the quality and shape of the corresponding meter by R, G and B color analysis of the extracted image; A fracture detection unit for generating a detection line next to the flank of the extracted image and determining whether a blade is detected in the detection line to determine fracture or dislocation; A controller for estimating the weight of the gauge from the weight table with respect to the measured total pixels, and determining the shape, quality, and fracture of the winter season from color signals of R, G, and B; A memory unit for storing a weight learning table and an inspection reference table set through data and learning necessary for operating the system, and temporarily storing weight estimation and quality inspection results; And a communication unit for connecting the data storage server and the image storage server to transmit the analyzed result data. The method of claim 1, And a weight learning table for weight determination and a learning table for inspection criteria are set from the total pixels of the image extracted from the acquired image. The method of claim 16, The weight learning table is a quantitative quality inspection system, characterized in that the weight of the real and the total amount of the pixel is set to match. Acquiring a stereoscopic image of the metrology in an image acquisition area of the metrology line; Extracting only an image of a metric, which is a target object, by separating the background from the acquired image; Extracting the extracted image by dividing the extracted image into parts; Estimating the weight of the meter through a set learning table by measuring the total amount of pixels of the extracted image; Analyzing the R, G, and B colors of the extracted image to detect quality inspection and shape defects; And generating the extracted weight data, the quality inspection and the shape defect result, and storing the result data and the image data. The method of claim 18, The image extraction of the metric system applies a threshold to process the pixel as a value above the threshold as '1', and to process the pixel below the threshold as a '0' to extract the image of the metric system and to paint the area to which the pixel is connected. A quantitative quality inspection method, characterized by recognizing a region consisting of the largest mass by adding the number of pixels as a target object. The method of claim 18, Wherein the segmentation is a differentiation along the outline of the image to obtain a deviation in the vertical while vertically, and the quantitative quality inspection method, characterized in that the portion in which the deviation is rapidly changed as the basis of the segmentation. The method of claim 18, The total pixel extraction is a quantitative quality inspection method comprising estimating the total volume by applying a certain amount of weights to the sum of the pixels excluding the wing and the neck which is a large deviation. The method of claim 1, The fracture analysis is a quantitative quality inspection method, characterized in that for generating a reference detection line and then determining the presence of the wing around the reference detection line. The method of claim 18, In the quality inspection for analyzing the R, G, B color, the quality of the farming quality inspection method characterized in that the problem is determined in the farming state, livelihood transport conditions, livestock processing process. The method of claim 18, The quality test for analyzing the R, G, and B colors is performed by circulating all the pixels of the recognized part of the image in the image while masking and matching the pixels corresponding to the color, size, and area, which are registered detection criteria, using the masked part. Names of defects: bruises, navel chickens, knee chickens, carnival 1, carnival 2, fowl, skin failure (grades A, B, C, D) and areas of failure (left leg, right leg, abdominal cavity, left) Abdomen, abdomen, right abdomen, left wing, chest, right wing, neck), color of the defect (R, G, B), the quality of the defect, characterized in that the detection method (mm). The method of claim 18, The quality test method of the quantitative quality characterized in that to use the data of the quality inspection result as feed back data for breeding management, transportation management, yard management process.

Images