The Design and Realization of Cherry Tomato Harvesting Robot Based on IOT

To improve the efficiency of harvesting cherry tomato and reduce its breakage rate, we design a harvesting robot base on image recognition and modular control. After image acquisition with IOT technology, the binary processing and expansion and corrosion processing of original image can effectively increase the fruit recognition rate. In addition, the use of fuzzy control technology processes the response error of manipulator. We test the performance of cherry tomato harvesting robot through harvesting experiment. The experimental results show that the harvesting efficiency significantly improves and the degree of crushing cherry tomato greatly reduces after using the cherry tomato harvesting robot.


INTRODUCTION
Cherry tomato is juicy with thin peel and easy to be damaged when manual harvesting.Besides, it must be harvested within a short time after they get ripe.Nowadays, labor cost in our country becomes higher and higher, while harvesting cherry tomato requires much labor.Also, its plants are short, so usually people need to bend down to pick the fruits.This means it will increase the burden of people.All of these promote the demands of machinery operation when harvesting cherry tomato.This research will design a robot, combining image capture technology, visual recognition technology and PLC control technology.It's expected that we can realize the mechanization and intelligentization of cherry tomato's harvesting process.

II. THE OVERALL DESIGN
The cherry tomato harvesting robot is mainly consists of four parts, which respectively are IOT vision sensor, four manipulators, controller and transmission device.The length, width and height of harvesting robot respectively are robot, 1M, 1.1M, 2.1M.Its three-dimensional vision sensor uses the image recognition technology to scan the cherry tomato and then recognize the fruit maturity, then using fuzzy control technology to determine the movement locus of cherry tomato harvesting; the overall design of cherry tomatoes harvesting robot shown in Figure 1.
Harvesting robot manipulator are composed of three levels of the system, which respectively are control system, stretching system and harvesting execution end The design of harvesting manipulator is shown in Figure 2.
Computer control system connects through the power bus, the control bus, the data bus and address bus and uses PLC control technology to conduct necessary segmentation process of image.PLC configuration is shown in Figure 3.

III. PLC CONTROL AND DESIGN
In the image recognition step, the priority is image segmentation.Supposing that the set is the entire image area and the image area of set has been divided into several sub-regions, in that way 1) , then is the connected area.
If , � is the empty set; means the union of all the subsets and also it can be understood as the collected original image area.
Meanwhile, in the procedure of image processing set 7 fuzzy sets as error differential In the operation, when

= = K K
can be worked out.According to the error feedback of PID parameters, the manipulator's rotary movement can be improved so that they can harvest more accurately and more efficiently.The working process as is shown in Figure 4.
The logic operations of control system as is shown in figure 5. IV.SYSTEM TEST Before putting the cherry tomato harvesting robot into work, we need test its performance under work status.Performance testing must do three aspects of test well: First is the robot identification of cherry tomatoes, second is the response of automation PLC harvesting control system, the last is the working efficiency of the parallel control.The cherry tomato images collected by image collector are shown in Figure 6.After the realization of the initial image collection, it needs binary image processing to convert into the electrical signal that can be recognized and processed by PLC.After processing binary image of Figure 7, Figure 8 is obtained.
Through the comparison test of image binary process whether using image dilation and erosion function, the conclusion is: the use of the dilation and erosion function can make the image contour of cherry tomato clearer, and the received image signal quality of PLC control system is better.The dilation and erosion effect figure of Figure 7 can be shown in Figure 8.
When testing the angle of harvesting robot, we find that if fuzzy control is not added, the system will have a large overshoot, which will make the system too unstable to output the robot angle.As shown in Figure 9.
By fuzzy control, the manipulator of cherry tomato harvesting robot can smoothly reach the specified angle to avoid overshoot situation, which can greatly reduce the breakage rate when harvesting cherry tomatoes.As shown in Figure10.
In order to highlight the contribution of this research robot to the cherry tomatoes harvesting work, we do harvesting broken rate statistics of this research cherry tomato harvesting robot and the traditional harvesting robot, of which make six comparison tests, as can be seen from test table 1 of statistical broken rate, the harvesting efficiency of this research robot significantly improves and the degree of crushing cherry tomato greatly reduces.11, the cherry tomato harvesting robot designed in this study can greatly shorten the harvesting operation time and improve the efficiency of harvesting robot when in harvesting work.

V. CONCLUSION
The visual recognition rate and localization precision of cherry tomato harvesting robot determines the accuracy of the harvest, this thesis implants binary image processing technology and dilation and erosion technology into the visual identification system of cherry tomato harvesting robot; It can improve the recognition rate of cherry tomatoes mature and decrease the harvesting error; To apply to PLC closed-loop feedback mechanism in the computer control system, which can improve the harvest accuracy of robot; It can reduce the crushing cherry tomatoes, make harvesting time significantly shorter and improve work efficiency of cherry tomato harvesting robot.

Figure 1 .Figure 2 .
Figure 1.The overall design diagram Cherry picking robot

K
the automatic closed-loop control mode of PLC and using tuning algorithm of PID parameters, set as the initial parameters of PID controller and set K as the sampling time, so

Figure 4 .Figure 5 .
Figure 4. PLC high-speed parallel design of Cherry picking manipulator

Figure 6 .
Figure 6.The cherry tomato images

TABLE I .
PAPER THE DESIGN AND REALIZATION OF CHERRY TOMATO HARVESTING ROBOT BASED ON IOTAt harvest time, we make comparison test between the cherry tomato harvesting robot designed in this study and the traditional harvesting robot, and the test results as shown in Figure BROKEN RATE TEST RESULTS % Figure 11.Picking time comparison test iJOE -Volume 12, Issue 12, 2016