CN106000912A - Citrus cyst foreign matter removing system based on machine vision guidance - Google Patents

Abstract

The invention discloses a citrus cyst foreign matter removing system based on machine vision guidance. The citrus cyst foreign matter removing system comprises a mechanical system, a vision measurement guidance system and a control system. The mechanical system comprises a frame, a delivery device and a removing device. The removing device comprises a horizontal movement platform and a foreign matter sucking device arranged at the upper end of the horizontal movement platform. The foreign matter sucking device comprises a sucking nozzle, a vacuum generator and a telescopic air cylinder which are connected in sequence. The sucking nozzle is connected with a foreign matter collecting box through the vacuum generator. The vision measurement guidance system comprises a light source, a camera unit and an image data acquisition card. The camera unit comprises a preprocessing camera arranged in a pre-detection area and a guidance camera arranged in a detecting and removing area. The control system comprises an industrial personal computer and a programmable logic controller (PLC). The control system controls starting of the delivery device, the light source and the camera unit, movement of the horizontal movement platform, and the action of the foreign matter sucking device. Compared with manual detection, the citrus cyst foreign matter removing system is easy to operate, low in omission ratio and high in efficiency and precision and has great market value.

Description

A system for removing foreign objects from citrus cysts based on machine vision guidance

technical field

The patent of the present invention relates to the field of fruit cyst foreign matter detection, in particular to a system for removing foreign matter from citrus cysts based on machine vision guidance.

Background technique

In recent years, my country's fruit juice beverage industry has developed rapidly, and the growth rate of output value has exceeded the growth rate of GDP. The good development prospects and the high degree of marketization of the entire industry have attracted international beverage giants to enter one after another, and the competition in the industry is also extremely fierce. However, in order to have a place in this industry, the quality of the product must be guaranteed. Citrus cysts are one of the agricultural products for making fruit juice. During the process of processing citrus to obtain cysts, it is inevitable to introduce a series of foreign substances such as capsule coating, orange core, hair and visible particles. Therefore, the process of removing cystic foreign matter must be carried out before the juice is canned, so as to ensure that the obtained juice is clean and delicious, and meets the consumption requirements of the public.

However, at present, there is no ready-made equipment in my country that can complete this process, and most companies' citrus cyst foreign body detection lines mainly use the working method of visual judgment and manual removal. This method has shortcomings such as low efficiency, high missed detection rate, and high cost, and the detection effect is also easily affected by factors such as the mood and physical state of employees. With the increasingly tight labor resources, these problems have become increasingly prominent. For this reason, an automatic detection system based on machine vision guidance is proposed to overcome the above problems.

Contents of the invention

The purpose of the patent of the present invention is to propose a system based on machine vision guidance to remove foreign matter from citrus cysts in view of the shortcomings of manual detection such as low efficiency, high missed detection rate and high cost. The system has the advantages of simple structure, simple operation, large detection range, high efficiency and high precision.

In order to achieve the above requirements, the patent of the present invention adopts the following technical solutions:

A system for removing foreign matter from citrus cysts based on machine vision guidance, characterized in that it includes a mechanical system, a visual measurement guidance system and a control system, the mechanical system includes a frame, a delivery device for transporting citrus cysts, and A rejecting device for removing foreign matter from citrus cysts, the rejecting device includes a moving translation platform and a foreign matter suction device arranged on the upper end of the moving translation platform, the moving translation platform can move horizontally above the conveying device, and the foreign matter suction device includes Suction nozzles, vacuum generators and retractable cylinders are connected in sequence, and the suction nozzles are also connected to the foreign matter collection box through the vacuum generators. The visual measurement guidance system includes a light source, a camera group, and an image data acquisition card. The camera group includes a pre-processing camera and a guiding camera. The conveying device is divided into a pre-detection area and a detection and rejection area according to the moving direction of citrus cysts. The pre-processing camera is arranged above the pre-detection area. Fixedly connected in the detection and rejection area, the control system includes an industrial computer and a PLC, the signals collected by the preprocessing camera and the guide camera are processed by the image data acquisition card and then connected to the industrial computer, and the industrial computer controls the conveying device through the PLC , the opening of the light source and camera group, the movement of the moving translation platform and the action of the foreign matter suction device.

As an improvement, the conveying device is arranged on the frame, and the conveying device includes driving rollers, driven rollers, a transmission belt connecting the driving rollers and driven rollers, and a drive motor driving the driving rollers, and the citrus cysts are transported The belt is transported, and the transmission belt is a transparent belt.

As an improvement, the coupling of the active roller is provided with a rotary encoder for measuring the rotational speed, and the detection signal of the rotary encoder is connected to the PLC, which can monitor the transmission speed of the active roller in real time, and control the conveying speed of the conveyor belt through the PLC .

As an improvement, the first and last ends of the conveying device are respectively provided with optical fiber sensors for detecting the citrus cysts entering and leaving the conveyor belt, and the signals of the optical fiber sensors are connected to the industrial computer to accurately judge whether the citrus cysts enter and leave the transmission belt. belt for accurate detection and removal of citrus cysts.

As an improvement, there are multiple rejecting devices, and the multiple rejecting devices are distributed in a linear array in the direction of the conveyor belt. The foreign matter suction device of each rejecting device is fixedly connected with a guide camera, and multiple rejecting devices work simultaneously. , to speed up the removal of foreign matter, effectively preventing missed detection and rejection.

The beneficial effects of the patent of the present invention are: the present invention is based on machine vision guidance to realize the system for removing foreign matter in citrus cysts. The visual measurement guidance system is used to collect pictures and transmit them to the industrial computer for analysis. Through the joint action of the moving translation platform and the cylinder, Guide the suction nozzle to remove foreign objects. The purpose of intelligent detection and removal of foreign matter is achieved, and the shortcomings of low efficiency and high cost of existing manual detection are overcome. Compared with manual detection, the operation is simple, the missed detection rate is low, the efficiency is high and the precision is high.

The speed of manual detection of citrus cysts is generally 0.6 square meters per minute, while the detection speed of the machine vision system can reach 5 square meters per minute. The speed of foreign matter removal by the machine vision system is 6 to 8 times that of manual labor, which greatly improves production efficiency .

Due to environmental and physiological reasons, manual detection cannot continuously detect and remove citrus cyst foreign bodies, but the use of machine vision system detection makes it possible. The production time of the equipment can be extended to the maximum extent, which improves the utilization rate of the equipment.

Manual detection is difficult to observe citrus cysts continuously due to visual fatigue. It is difficult to guarantee the correct detection rate by manual work. The detection accuracy of the machine vision system is high, which greatly improves the product qualification rate and detection quality.

Description of drawings

Fig. 1 is a structural representation of the present invention;

Fig. 2 is a block diagram of the composition of the system of the present invention;

Fig. 3 is a work flowchart of the present invention;

Fig. 4 is an image-based visual servo system;

Figure 5 is the visual servoing control flow.

In Figure 1, 1 universal wheel, 2 frame, 3 driving roller, 4 conveyor belt, 5 preprocessing camera, 6 guiding camera, 7, cylinder, 8 vacuum generator, 9 suction nozzle, 10 foreign body collection box, 11 industrial computer , 12 trachea, 13 motion translation stage, 14 guide rail, 15 image data acquisition card, 16 IEEE1394B data line, 17 PLC, 18 solenoid valves, 19 drivers, 20 driven rollers.

detailed description

In order to facilitate those of ordinary skill in the art to understand and implement the patent of the present invention, the patent of the present invention will be described in further detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the implementation examples described here are only used to illustrate and explain the patent of the present invention, and It is not used to limit the invention patent.

Below in conjunction with embodiment and accompanying drawing, the patent of the present invention is described further.

The invention is applied to the removal of foreign matter in citrus cysts. It mainly focuses on foreign matter such as capsule coat, orange nucleus, hair, and fiber mixed in the cyst, which involves technologies in visual recognition, visual guidance, image processing, mechanical transmission, and sensors. A new scheme for removing foreign bodies from citrus cysts by using machine vision guidance technology was proposed.

The purpose of this system is to replace the manual removal of foreign matter in citrus cysts on the production line, because the detection and removal of foreign bodies in citrus cysts are currently performed manually. This system achieves the purpose of intelligent detection and removal of foreign bodies, overcoming the existing artificial The shortcomings of low detection efficiency and high cost. Compared with manual detection, the operation is simple, the missed detection rate is low, the efficiency is high and the precision is high.

As shown in Figure 1, a citrus cyst foreign body removal system based on machine vision guidance includes three main parts: mechanical system, visual measurement guidance system and control system. The mechanical part mainly includes frame 2, conveying device and rejecting device. The motion translation platform 13 is responsible for carrying the actuators such as the cylinder 7 and the suction nozzle 9 to move above the foreign matter, and the cylinder 7 is responsible for driving the suction nozzle 9 to move up and down and absorbing foreign matter; the visual measurement guidance system is composed of a light source, a camera group, and image data collection. Composed of card 15, it is responsible for the identification of foreign matter such as capsule coat, orange core, hair, fiber, etc. mixed in the cyst; the control system controls the opening of the conveying device, light source and camera group, the movement of the moving translation platform 13, and the action of the foreign matter suction device .

The conveying device is divided into a pre-detection area and a detection and rejection area according to the direction of movement of citrus cysts. The camera group includes a pre-processing camera 5 installed in the pre-detection area and a pair of guide cameras 6 installed in the detection and elimination area. The camera 5 mainly detects whether there are foreign objects in the citrus cysts, and guides the camera 6 to track and locate the foreign objects in real time, and cooperates with the rejecting device to remove the foreign objects;

The image data acquisition card 15 is respectively connected to the camera group by a data cable, and the image data acquisition card 15 is connected to the image processing unit of the industrial computer 11; the optical fiber sensor is a through-beam optical fiber sensor, symmetrically arranged on the transmission belt On the sensor brackets on both sides of the sensor, the output signal of the fiber optic sensor is introduced into the PLC 17. The PLC 17 is also connected to the light source and the camera group.

The cameras of the camera group are all installed on the support with special fixtures, according to the size of the citrus cysts and the tiled area, the focal length of the lens is selected, and the shooting angle, shooting distance, aperture size, and exposure time of the shooting camera are adjusted, so as to obtain Clear shooting images; according to the design, this embodiment selects the GuppyPro 125c color CCD camera of German AVT Company for use. Industrial fixed focus lens. Adjust the height of the camera (that is, the object distance) so that the field of view of the camera is a 420*313mm area; at this time, the pixel resolution of the system is 0.33mm/pixel; the theoretical maximum acquisition speed is 9.7m/s; considering the image processing time and mechanical Action time, the system processing speed is not lower than 500mm/s.

Described delivery device comprises driving roller 3 and driven roller 20, the three-phase asynchronous motor that links to each other with driving roller 3, and the transmission belt 4 between driving roller 3 and driven roller 20; Described transmission belt 4 is a transparent transmission belt , and the material is smooth, allowing citrus cysts to be laid flat on the conveyor belt 4. The inner diameter is about 50 cm wide and the fastest flow rate is 10m/min; the driving roller 3 is equipped with a rotary encoder through a coupling to monitor the conveyor belt in real time The speed, the output signal of the rotary encoder is introduced into the programmable controller PLC 17.

The rejecting device includes a moving translation table 13, a cylinder 7, an electromagnetic valve 18, a vacuum generator 8, a suction nozzle 9 and a foreign matter collection box 10, and the suction nozzle 9 is connected under the vacuum generator 8, and relies on the vacuum generator 8 to provide The foreign matter is sucked by power, and the foreign matter collection box 10 is connected to one side of the vacuum generator 8 through an air pipe, and the electromagnetic valve 18 is used to control the movement of the cylinder 7 .

The rejecting devices are distributed in a linear array on the conveyor belt 4, and their range of movement in the horizontal direction is limited within a certain range, and each rejecting device does not affect each other during operation.

The overall block diagram of the system for removing foreign bodies from citrus cysts based on machine vision guidance is shown in Figure 1. When the cyst moves with the conveyor belt 4, it passes through the pre-detection area and the detection and rejection area successively. In the pre-detection area, after the image is collected by the pre-processing camera 5, it is transmitted to the industrial computer 11 through the IEEE1394B data line 16, and the industrial computer 11 processes and analyzes the current image to extract foreign object features and corresponding coordinate information. The PLC 17 drives the moving translation platform 13 to a corresponding position according to the lateral coordinates of the foreign matter, and waits for the arrival of the foreign matter. When a foreign object enters the field of view of the guiding camera 6, the guiding camera 6 starts to track and locate the foreign object in real time. After the foreign object is locked, the guiding motion translation table 13 moves directly above the foreign object, the cylinder 7 is extended, and the vacuum generator 8 generates negative pressure. , The suction nozzle 9 removes foreign matter.

The specific workflow of the system is shown in Figure 3: the system starts to work, and each part starts. The rotary encoder measures the real-time speed of the conveyor belt 4 and the driving roller 3, and the PLC 17 compares it with the set value. If it does not match, the PLC 17 sends an adjustment signal, and the frequency converter adjusts the motor speed to adjust the linear speed of the conveyor belt; When the fiber optic sensor next to the conveyor belt detects that the citrus cysts enter the shooting section, a level signal is generated to trigger the strobe of the light source and the camera to capture images of the citrus cysts.

From entering the field of view of the pre-processing camera 5 to the arrival of the foreign matter in the collection box 10 as a working cycle. The specific process of each work cycle is that the image captured by the preprocessing camera 5 is transferred to the industrial computer 11 through the image data acquisition card 15, and the image processing software in the industrial computer 11 performs a series of image processing, and the industrial computer 11 detects the field of view. All the foreign objects inside, and input the coordinates of the foreign objects into the database. After optimizing the data in the database, the motion translation platform 13 starts to move, ready to welcome the arrival of foreign objects. When arriving at the shooting area of the guiding camera 6, the guiding camera 6 continuously collects images and transmits the images to the industrial computer 11 in real time. The industrial computer 11 detects whether there is a foreign object in the field of view. If not, the working cycle ends. If there is a foreign object, an instruction is sent to make the moving translation platform 13 move along with the foreign object to lock the target foreign object. At this time, the electromagnetic valve 18 controls the action of the cylinder 7, and the cylinder 7 drives the suction mechanism to extend, and the suction nozzle 9 generates negative pressure through the vacuum generator 8 to suck foreign matter. After the foreign matter is removed, each mechanism resets. This loops until there is no foreign object in the field of view of the guiding camera 6 . Then enter the next working cycle.

A few points to elaborate on:

In the detection area, the adopted image-based visual servo system is shown in Figure 4. Firstly, the guided camera is used to collect images of citrus cysts in real time, and the corresponding image coordinates and motion information are extracted. Furthermore, image-based visual servoing technology is used to drive the movement of the culling device, so that the target always appears in the field of view of the guiding camera 6, so as to achieve the purpose of visually locking the target. In this way, the tracking and absolute positioning of foreign bodies in citrus cysts can be realized.

The purpose of feature teaching is to acquire the image of the foreign object in the guide camera 6 when the suction nozzle 9 is in the desired pose, and use it as the desired feature of visual servoing. When the system is in the working mode, since the current relative pose of the foreign object and the suction nozzle 9 usually deviates from the expected relative pose, the image acquired by the camera is also different from the expected image features. Taking the difference as the input of the controller, the controller updates the image Jacobian matrix and its inverse matrix in real time, and calculates the output control quantity, so as to control the movement of the cylinder through the solenoid valve 18 . Recursively, until the deviation of the image features is less than the allowable error range, and finally reach the precise operating position.

The overall workflow is shown in Figure 5. The purpose of the tentative exercise is to establish the initial Jacobian matrix.

This solution does not need to estimate the pose of the cystic foreign body in the Cartesian coordinate system, which not only reduces the calculation delay, but also overcomes the influence of the calibration error of the camera and the error of the nozzle position sensor on the positioning accuracy, and even does not require calibration. The calculated image Jacobian matrix and its inverse matrix link the change of image feature parameters with the change of nozzle pose. Its essence is to describe and control the servo task in the image feature parameter space. Compared with the position-based servo method, this method has better positioning and tracking accuracy.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection of the present invention. within range.

Claims (5)

1. one kind realizes Citrus cyst foreign body eliminating system based on machine vision guiding, it is characterised in that: comprise mechanical system, regard Feel and measure guidance system and control system, described mechanical system include framework (2), for carry Citrus cyst conveyer device, And for rejecting the device for eliminating of Citrus cyst foreign body, described device for eliminating includes motion translation platform (13) and is located at motion and puts down The foreign body suction means of moving stage (13) upper end, described motion translation platform (13) can the most just move horizontally, described Foreign body suction means includes suction nozzle (9), vacuum generator (8) and telescopic cylinder (7), the described suction nozzle being sequentially connected (9) be connected with foreign matter collection case (10) also by vacuum generator (8), described vision measurement guidance system include light source, Phase unit, Image Data Acquisition Card (15), described phase unit includes pretreatment camera (5) and guides camera (6), described defeated Sending device to be divided into pre-detection district and detection to reject district according to the Citrus cyst direction of motion, described pretreatment camera (5) is located at pre-detection Above district, described guiding camera (6) and foreign body suction means are fixedly linked and are located at detection rejecting district, and described control system includes work Control machine (11) and PLC (17), described pretreatment camera (5) and the signal guiding camera (6) to gather pass through view data Capture card (15) process after access industrial computer (11), described industrial computer (11) by PLC (17) control conveyer device, Light source and the unlatching of phase unit, the motion of motion translation platform (13) and the action of foreign body suction means.

2. a kind of guiding based on machine vision realizes Citrus cyst foreign body eliminating system, it is characterised in that: Described conveyer device is located on framework (2), and described conveyer device includes active roller (3), return idler (20), connects master Move roller (3) and the transmission belt (4) of return idler (20) and drive the driving motor of active roller (3), described Citrus chachiensis Hort. Fructus Citri tangerinae cyst is carried by transmission belt (4), and described transmission belt (4) is transparent belt.

3. a kind of guiding based on machine vision realizes Citrus cyst foreign body eliminating system, it is characterised in that: The shaft coupling of described active roller (3) is provided with the rotary encoder for measuring rotating speed, and described rotary encoder detection signal connects Enter PLC (17).

4. a kind of guiding based on machine vision realizes Citrus cyst foreign body eliminating system, it is characterised in that: Described conveyer device head and end is respectively provided with detection Citrus cyst and enters and leaves the Fibre Optical Sensor of transmission belt (4), institute State Fibre Optical Sensor signal and access industrial computer (11).

5. a kind of guiding based on machine vision realizes Citrus cyst foreign body eliminating system, it is characterised in that: Described device for eliminating has multiple, the plurality of device for eliminating linearly array distribution on transmission belt (4) direction, described each The foreign body suction means of device for eliminating has all been fixedly linked guiding camera (6).