CN1394699A - Fruit quality real time detection and grading robot system - Google Patents

Abstract

The invention discloses a fruit quality real-time detection and grading robot system. It is composed of fruit conveying turning parts, computer vision recognition parts, and automatic grading parts. The double-cone rollers of the fruit conveying and turning parts make the fruits automatically arranged in a single row, and while being conveyed forward at a certain speed, they rotate evenly around the small horizontal axis, so as to ensure that the computer vision recognition part can obtain the quality information of the entire surface of the fruit . Through the recognition of computer vision recognition components, the detection of all appearance quality indicators such as the shape, size, color, skin smoothness, fruit surface defects and damages of the fruit is completed at the same time, the grade of each fruit is comprehensively judged, and its location information is determined. The control module of the computer vision recognition part transmits the instruction to the automatic grading part, and controls the fruit to automatically fall into the fruit collection box at the corresponding grading port. It can quickly and effectively realize the real-time detection and grading of the dynamic fruit on the production line, and improve the quality of the fruit. The automation level of quality inspection and grading.

Description

Fruit quality real-time detection and grading robot system

Technical field

The invention relates to a robot system for real-time detection and classification according to the characteristics of objects, in particular to a robot system for real-time detection and classification of spherical fruits or agricultural products.

Background technique

my country is a large agricultural country, but its agricultural output value is very low. One of the main reasons is that the level of post-production commercialization of agricultural products in our country is too low, which greatly hinders the development of our national economy and the increase of farmers' income. In developed countries, the vast majority of agricultural output value is created by the post-harvest processing of agricultural by-products. For example, 75% of agricultural products in the United States are processed. However, the level of post-production treatment of agricultural and sideline products in our country is very low. For example, fruit production occupies a large proportion in the production of agricultural products in our country. Since 1993, the total amount of fruits in our country has leapt to the first in the world, and there are many types and varieties. At the same time, fruit is also an important foreign trade export material in my country. However, due to insufficient commercialization of post-harvest fruits, the products are mixed and graded, good and bad are intermingled, and export fruits cannot be placed on foreign high-end shelves, product reputation is low, and competitiveness in the international market is weak. The backwardness of the means of selection. After foreign agricultural products are automatically graded and packaged according to size, shape, color, damage and defects, their commodity value is greatly improved. At present, most of the agricultural products in our country are directly put on the market without treatment, and the price is naturally very low. A small amount of post-production commercial treatment is usually done by people, which requires a lot of labor and labor intensity. At the same time, this subjective evaluation is affected by factors such as personal vision, color discrimination, emotion, fatigue, light, etc., so the efficiency is low and the accuracy is poor. Although size grading can achieve mechanized grading based on size and weight, which reduces some labor intensity, it cannot complete the detection and grading of fruit shape, color, fruit surface defects and damage.

It is self-evident that it is self-evident to use computer vision technology and robot technology to replace people in the quality inspection and grading of agricultural products. First of all, it can eliminate the interference of human subjective factors and avoid the detection results that vary from person to person; in addition, it can complete tasks that are difficult for humans or mechanical or photoelectric sorting machines, such as the calculation of damaged area and stained area. It can not only improve the precision, but also liberate people from heavy labor. In recent years, with the rapid development of computer vision technology, this technology has begun to be applied to the detection and classification of fruits, but currently it is mainly focused on the processing of static fruits. Although there are foreign countries that can detect and classify one, two or at most three indicators of the appearance quality indicators such as fruit shape, size, color, skin smoothness, fruit surface defects and damage, there is no equipment that can simultaneously The equipment that completes the detection of all the above appearance quality indicators; moreover, the fruit conveying system in the existing equipment cannot guarantee that the fruits are automatically conveyed in a single row and turned over quickly and evenly, and can only enable the computer vision system to obtain image information on the surface of the detected fruit. The missed detection rate of fruit surface damage or defect is high; in addition, real-time detection and grading equipment still has problems such as low detection accuracy and slow speed. In China, similar testing and grading equipment is still blank.

Contents of the invention

The invention provides a robot system for real-time detection and grading of fruit quality, which can realize fast and uniform turning over of fruits and automatic single-row conveyance, and present the fruits on the computer vision recognition part in a suitable and constantly changing position and posture In the field of view, the computer vision recognition component can accurately, effectively and comprehensively obtain the quality characteristic information of the detected object, and can simultaneously complete all appearance quality of the fruit such as shape, size, color, skin smoothness, fruit surface defects and damage The detection of indicators realizes the real-time detection and grading of fruits.

The technical scheme that the present invention adopts is as follows:

1) Fruit conveying turning parts: including double-cone rollers, conveyor chains, friction belts, conveyor sprockets, slope plates, and horizontal small shafts; double-cone rollers are evenly installed on the conveyor chain through the horizontal small shafts, and the When the sprocket rotates, the double-cone rollers can move forward with the conveyor chain. The friction belt installed under the double-cone rollers is driven by another motor, and the double-cone rollers will rotate around the small horizontal axis on the friction belt. , there are ribbed slope plates on both sides of the double cone roller, and the lowest end of the slope plate is slightly lower than the double cone roller;

2) Computer vision identification components: including a camera, a light box, a computer, a control module, and two position sensors. The two position sensors are installed on the fruit conveying turning part, and are respectively connected to the image acquisition card and the control module in the computer through lines. And provide the position information of each fruit, the light box is installed above the fruit conveying and turning parts, the camera is installed in the light box, connected with the image acquisition card in the computer through the line, the fruit position information provided by the position sensor triggers the image in the computer The acquisition card collects dynamic fruit images, the computer is connected to the control module through the parallel port, and the control module sends signals to control the corresponding grading executive mechanism and grading drive mechanism to send the fruit to the corresponding collection box;

3) Fruit automatic grading components: including grading sprockets, grading chains, hopper shafts, grading hoppers, grading drive mechanisms, fruit collection boxes of different levels, fruit falling slides, and fruit grading output mechanisms; the grading hoppers pass through the hopper shafts evenly The ground is installed on the grading chain, the grading sprocket rotates, the grading hopper can move forward with the grading chain, the control module sends a signal to control the grading drive mechanism, the grading hopper loses stability, and the fruit is graded and output along the slideway installed under the grading drive mechanism Body to fruit collection box.

The double-cone roller fruit conveying and turning device of the fruit conveying and turning part can automatically convey the fruits forward in a single row at a certain speed, and can make the fruits rotate freely around the small horizontal axis, so as to ensure that the entire surface of the fruits can be detected. Setting the computer vision recognition component above the fruit conveying and turning component can ensure accurate, effective and comprehensive acquisition of the image information of the detected object, and extract quality feature information therefrom. Through the intelligent recognition of computer vision recognition components, the detection of all appearance quality indicators such as fruit shape, size, color, skin smoothness, fruit surface defects and damage is completed at the same time, and the grade of each fruit is comprehensively judged, and the quality of each fruit is determined. For the location information, the control module of the computer vision recognition component transmits instructions to the automatic grading component, and the automatic grading component completes the fruit grading according to the fruit grading standard.

The beneficial effect of the present invention is that effective image information can be extracted from fast-moving fruit groups at the same time, corrected and analyzed, and the shape, size, color, skin smoothness, and fruit shape of dynamic fruits on the production line can be quickly and effectively completed. The detection of all appearance quality indicators such as surface defects and damages, and the automatic grading components are used to achieve grading, which improves the automation level of fruit quality detection and grading. After being applied to actual production, it can improve the competitiveness of our country's fruits in the international market.

Description of drawings

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

Fig. 2 is a schematic diagram of the fruit conveying turning part of the present invention;

Fig. 3 is a sectional view of Fig. 2;

Fig. 4 is a schematic structural view of the automatic grading component of the present invention;

Figure 5 is a top view of Figure 4;

Fig. 6 is a schematic diagram of hierarchical output;

Figure 7 is a structural block diagram of the control module.

In the figure: 1. Fruit conveying turning part 2, computer vision recognition part 3, automatic grading part 4, camera 5, light box 6, computer 7, control module 8, grading actuator 9, position sensor 10, double cone roller 11. Fruit 12, conveyor chain 13, friction belt 14, conveyor sprocket 15, slope plate 16, horizontal small shaft 17, grading sprocket 18, grading chain 19, hopper shaft 20, grading hopper 21, grading drive mechanism 22, Fruit collection box 23, drop slideway 24, classification output mechanism

Detailed ways

As shown in FIG. 1 , the present invention consists of a fruit conveying turning part 1 , a computer vision recognition part 2 and a fruit grading part 3 .

As shown in Fig. 1, Fig. 2 and Fig. 3, the fruit conveying overturning part 1: comprises a double-cone roller 10, a conveying chain 12, a friction belt 13, a conveying sprocket 14, a slope plate 15, and a small horizontal shaft 16; The tapered rollers 10 are evenly installed on the conveying chain 12 through the horizontal small shaft 16, and the conveying sprocket 14 rotates. The friction belt 13 is driven by another motor. When there is a speed difference between the two, the double-cone roller 10 will rotate around the small horizontal axis 16 on the friction belt 13. Ribbed rollers are provided on both sides of the double-cone roller 10. The slope plate 15, the lowest end of the slope plate 15 is slightly higher than the double tapered roller 10. After the fruit 11 enters the device, due to the inclination of the double-cone roller 10 itself, the fruit can automatically enter the middle of each pair of double-cone rollers 10 and contact the four cones at the same time. It is to ensure that during the conveying process of fruits, the fruits can automatically enter each pair of double-cone rollers 10 in a row, and under the action of a pair of double-cone rollers 10, the fruits can be transported forward and turned over at the same time. The computer vision recognition system, the conveying and turning speed of the fruit conveying and turning parts can be adjusted according to the needs of the production line work.

As shown in Figure 1, the computer vision identification part 2: comprises camera 4, light box 5, computer 6, control module 7, two position sensors 9, two position sensors 9 are installed on the fruit conveying turning part 1, respectively through the circuit Link to each other with the image acquisition card and control module 7 in the computer 6, and provide the position information of each fruit, the light box 5 is installed on the top of the fruit conveying turning part 1, and the camera 4 is installed in the light box 5, through the circuit and the computer 6 The image acquisition card is connected to each other, and the fruit position information provided by the position sensor triggers the image acquisition card in the computer to collect dynamic fruit images. The computer 6 is connected to the control module 7 through the parallel port, and the control module 7 sends signals to control the corresponding hierarchical actuators 8 and 2. Grading drive mechanism 21 sends fruit 11 in the corresponding collection box 22. Image analysis and processing software suitable for fast processing of dynamic images is installed in the computer 6 . The illumination box 4 makes the illumination in the entire field of view uniform by selecting an appropriate light source spectrum and spatial position, and can be adjusted according to different identification objects. The camera 4 is installed in the light box 5, and by adjusting the position, it can obtain the image information of a plurality of fruits, and the image processing and analysis software analyzes the shape, size, color, surface smoothness, surface defects, Extract, analyze and judge the appearance quality characteristics such as damage, and determine the grade of the fruit according to the national standard classification. Position sensor 9 is used to determine the information of fruit position.

As shown in Fig. 4, Fig. 5, Fig. 6, fruit automatic grading part 3: comprise grading sprocket wheel 17, grading chain 18, hopper shaft 19, grading hopper 20, grading drive mechanism 21, fruit collection box 22 of different levels, fruit The classification output mechanism 24 of falling slideway 23, fruit; When the fruit with position information was transported to the position of the fruit collection box 22 of the corresponding level, the position signal of the grading drive mechanism 21 sent by the control module 7 of the computer vision recognition part drove the grading drive mechanism 21, so that the grading hopper 20 Loss of stability, the fruit falls along the falling chute 23 at the corresponding grading port, and is transported in the fruit collection box 22 by the grading output mechanism 24, and realizes the grading of the fruit.

Figure 7 is a schematic diagram of the control module 7, the control module 7 is composed of Advantech embedded motherboard PCM-3346N, electronic disk, touch screen and 48-channel input and output module PCM-3724 based on PC104 expansion bus. The electronic disk is connected to the main board through the special interface on the main board, the touch screen is connected to the main board through the RS232C serial port, the digital input and output module is connected to the main board through the PC1104 bus, and the fruit grade and position information are connected to the main board through the serial port RS232C.

PCM-3346N adopts low power consumption GX1-300MHz CPU, working temperature: -20℃～+85℃, supports CompactFlash, and has watchdog circuit, stable performance, suitable for industrial control needs. The electronic disk adopts the 128M CF card of Sandisk Company, which is connected with the main board through the special interface on the input main board PCM-3346N, and is used to store the embedded operating system and application software. The touch screen is used to display the working status of the system, set working parameters, input start and stop and other control commands, adopt AccuTouch company 104R/IP65 type, and connect with the main board PCM-3346N through the RS232C serial interface. The digital input and output module PCM-3724 has an interrupt trigger function, and is connected to the PCM-3346N main board through the PC/104 bus, and is used to collect equipment working status information and control the grading actuator in the automatic grading component.

The main function of the control module 7 is to receive the position information and grade information of the fruit, and send a control signal to the grading actuator 8 . After the grading starts, when a fruit passes through the position sensor 9 and the computer vision identification part 2, the position information and the grade information of the fruit are obtained. The position and grade information of the fruit enter the PCM-3346N main board through the serial interface RS232C, when the When the fruit reaches the position of the corresponding grading port, the PCM-3346N main board sends a signal to the 48-channel input and output module PCM-3724 through the PC104 bus, and the 48-channel input and output module PCM-3724 controls the corresponding grading actuator 8 and the grading drive mechanism 21, so that The fruit falls into the fruit collection box 22 of the corresponding level. In addition, the control module also has functions such as fruit grading process control, equipment working status monitoring in the grading system, and quick reset of the control system.

For different fruits, different detection speeds are required. The fruit conveying and turning system can be adjusted synchronously with the automatic grading system, and the image sampling frequency of the computer vision recognition component is also automatically adjusted according to the speed change.

Claims (2)

1. A robot system for real-time detection and grading of fruit quality, characterized in that it includes: 1) Fruit conveying turning parts (1): including double-tapered rollers (10), conveying chains (12), friction belts (13), conveying sprockets (14), sloped plates (15), horizontal small shafts ( 16); double-tapered rollers (10) are evenly contained on the conveyor chain (12) by the horizontal small shaft (16), and the conveyor sprocket (14) rotates, and the double-tapered rollers (10) can follow the conveyor chain (16). 12) Moving forward, the friction belt (13) installed under the double-tapered roller (10) is driven by another motor, and the double-tapered roller (10) will wind around the small horizontal shaft on the friction belt (13). (16) rotate, double-cone type roller (10) both sides are provided with ribbed slope plate (15), and the lowest end of slope-shaped plate (15) is slightly lower than double-cone type roller (10); 2) Computer vision recognition component (2): including camera (4), light box (5), computer (6), control module (7), two position sensors (9); two position sensors (9) are installed on The fruit conveying turning part (1) is connected with the image acquisition card and the control module (7) in the computer (6) respectively through lines, and the position information of each fruit is provided, and the light box (5) is installed on the fruit conveying turning part (1) top, camera (4) is installed in the light box (5), links to each other with the image acquisition card in the computer (6) through the circuit, the fruit position information provided by the position sensor (9) triggers in the computer (6) The image acquisition card collects dynamic fruit images, and the computer (6) is connected to the control module (7) through the parallel port, and the control module (7) sends a signal to control the corresponding grading executive mechanism (8) and the grading drive mechanism (21) to drive the fruit (11 ) into the corresponding collection box (22); 3) Automatic grading components (3): including grading sprockets (17), grading chains (18), hopper shafts (19), grading hoppers (20), grading drive mechanisms (21), fruit collection boxes of different levels (22 ), the grading output mechanism (24) of the whereabouts slideway (23) of fruit, fruit; Grading hopper (20) is contained on the grading chain (18) equably by hopper shaft (19), and grading sprocket wheel (17) rotates, The grading hopper (20) can move forward with the grading chain (18), the control module (7) sends a signal to control the grading drive mechanism (21), the grading hopper (20) loses stability, and the fruit (11) is installed on the grading drive mechanism ( 21) down the slideway (23) to the fruit collection box (22) through the classification output mechanism (24). 2. fruit quality real-time detection and grading robot system according to claim 1, it is characterized in that said control module (7) is the CF card of Sandisk company, touch screen and based on 48-channel input and output module PCM-3724 of PC104 expansion bus; the electronic disk is connected to the main board through the special interface on the main board, the touch screen is connected to the main board through the RS232C serial port, and the digital input and output module is connected to the main board through the PC/104 bus. And position information is connected to the main board through the serial connection RS232C.

Images