• The Journal of Korea Robotics Society
• /
• v.15 no.2
• /
• pp.107-114
• /
• 2020

In robotic harvesting, a gripper to manipulate the fruits needs to be attached to the robot system. We proposed a flexible robot gripper that can actively respond to the shape of an object such as fruits in the previous work. However, we found that there is a possibility of not being reliably gripped when the object slides during contact with a finger. In this paper, the improved gripper design is proposed to fundamentally solve the problems of the previous gripper. The position of the finger and the maximum closed position are changed, and the design improvement is performed to increase the grip stability by changing the installation angle of the link portion of the finger. Based on the improved design, a modified gripper is fabricated by 3-D printing, and then gripping experiments are performed on spherical object and fruit model object. It is shown that the gripper can stably grip the objects without excessive bending of the finger link of the gripper. The contact pressure between the finger and the surface of the object is measured, and it is verified that it is a sufficiently small pressure that does not cause damage to the fruit. Therefore, the proposed gripper is expected to be successfully applied in harvesting.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.2
• /
• pp.421-441
• /
• 2021

Fruit detection in orchards is one of the most crucial tasks for designing the visual system of an automated harvesting robot. It is the first and foremost tool employed for tasks such as sorting, grading, harvesting, disease control, and yield estimation, etc. Efficient visual systems are crucial for designing an automated robot. However, conventional fruit detection methods always a trade-off with accuracy, real-time response, and extensibility. Therefore, an improved method is proposed based on coarse-to-fine multitask cascaded convolutional networks (MTCNN) with three aspects to enable the practical application. First, the architecture of Fruit-MTCNN was improved to increase its power to discriminate between objects and their backgrounds. Then, with a few manual labels and operations, synthetic images and labels were generated to increase the diversity and the number of image samples. Further, through the online hard example mining (OHEM) strategy during training, the detector retrained hard examples. Finally, the improved detector was tested for its performance that proved superior in predicted accuracy and retaining good performances on portability with the low time cost. Based on performance, it was concluded that the detector could be applied practically in the actual orchard environment.

• The Journal of Korea Robotics Society
• /
• v.16 no.1
• /
• pp.49-55
• /
• 2021

Recently, a smart farm technology is drawing attention as an alternative to the decline of farm labor population problems due to the aging society. Especially, there is an increasing demand for automatic harvesting system that can be commercialized in the market. Pre-harvest crop detection is the most important issue for the harvesting robot system in a real-world environment. In this paper, we proposed a real-time tomato instance tracking algorithm by using deep learning and probability models. In general, It is hard to keep track of the same tomato instance between successive frames, because the tomato growing environment is disturbed by the change of lighting condition and a background clutter without a stochastic approach. Therefore, this work suggests that individual tomato object detection for each frame is conducted by YOLOv3 model, and the continuous instance tracking between frames is performed by Kalman filter and probability model. We have verified the performance of the proposed method, an experiment was shown a good result in real-world test data.

• Journal of Drive and Control
• /
• v.21 no.2
• /
• pp.65-71
• /
• 2024

In this study, we suggested a key-point detection method for robot harvesting of oriental melon. Our suggested method could be used to detect the detachment part and major composition of oriental melon. We defined four points (harvesting point, calyx, center, bottom) based on tomato with characteristics similar to those of oriental melon. The evaluation of estimated key-points was conducted by pixel error and PDK (percentage of detected key-point) index. Results showed that the average pixel error was 18.26 ± 16.62 for the x coordinate and 17.74 ± 18.07 for the y coordinate. Considering the resolution of raw images, these pixel errors were not expected to have a serious impact. The PDK score was found to be 89.5% PDK@0.5 on average. It was possible to estimate oriental melon specific key-point. As a result of this research, we believe that the proposed method can contribute to the application of harvesting robot system.

• Smart Media Journal
• /
• v.12 no.10
• /
• pp.19-28
• /
• 2023

In the fruit harvesting field, interest in automatic robot harvesting is increasing due to various seasonality and rising harvesting costs. Accurate apple detection is a difficult problem in complex orchard environments with changes in light, vibrations caused by wind, and occlusion of leaves and branches. In this paper, we introduce a dataset and an adaptive heatmap regression model that are advantageous for robot automatic apple harvesting. The apple dataset was labeled with not only the apple location but also the visibility. We propose a method to detect the center point of an apple using an adaptive heatmap regression model that adjusts the Gaussian shape according to visibility. The experimental results showed that the performance of the proposed method was applicable to apple harvesting robots, with MAP@K of 0.9809 and 0.9801 when K=5 and K=10, respectively.

• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 1996.06c
• /
• pp.638-647
• /
• 1996

Robot manipulators for harvesting fruits must be controlled to track the desired path of end-effector to avoid obstacles under the consideration of collision free area and safety path. This paper presents a robot path control algorithm to secure a collision free area with the recognition of work environments. The flexible space, which does not damage fruits or branches of tree due to their flexibility and physical properties , extends the workspace. Now the task is to control robot path in the extended workspace with the consideration of collision avoidance and velocity limitation at the time of collision concurrently. The feasibility and effectiveness of the new algorithm for redundant manipulators were tested through simulations of a redundant manipulator for different joint velocities.

• Journal of Biosystems Engineering
• /
• v.13 no.2
• /
• pp.9-17
• /
• 1988

This study was carried out to develop an agricultural robot for fruit harvesting. As the first step an experimental manipulator and its control system were constructed. The articulated manipulator driven by DC motors has 3 degrees-of-freedom. The manipulator has a gripper adequate for fruit harvesting and an upper arm which forms a kind of guiding channel so thai harvested fruit can pass through. Point-to-point control of joints are accomplished by a digital control system with a PID controller which consists of optical shaft encoders, power amplifiers using PWM, a microcomputer and a software. The microcomputer also computes the positions of manipulator and sequence of motions. The motion of the manipulator was to slow and rough that it would need further improvement.

• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 1993.10a
• /
• pp.840-849
• /
• 1993

In this paper, it is reported that manipulator and hand-required for harvesting tomato were studied. At first, basic physical properties of tomato plant were investigated such as position of fruit, length of stems and leaves, width between ridges and son on . Secondly , basic mechanism of articulate manipulators with 5 to 7 degree of freedom were investigated by using evaluation indexes such as operational space, measure of manipulatability , posture diversity and so on. From the results, an articulate manipulator with 7 degrees of freedom was selected and the manipulator was manufactured as a trial according to the mechanism. Thirdly , physical properties about fruit and peduncle of tomato were also researched such as diameter, length , picking force and so on. Based on the properties , tomato harvesting hand with absorptive pad were also made as a trial. Finally, after the hand was attached to the manipulator, harvesting experiment was done in greenhouse . It was observed th t the robot could harvest satisfactorily , not only since the robot adapted to physical properties of tomato plant was manufactured but also since phyllotaxis of tomatoes was so methodical that all fruit clusters emerged in the same direction.

• Journal of Korea Multimedia Society
• /
• v.15 no.11
• /
• pp.1273-1283
• /
• 2012

In this paper, vision-based positioning algorithm for melon harvesting robot is presented. RGB value of the input image was converted into HSI value then, melon area was extracted after performing the binarization using HUE value. After morphological filtering was applied to remove noise, outermost boundary points were obtained using border following and convex hull method. Elliptical fitting for melons was perform by the RANSAC algorithm, the center point of ellipse, the length of the short and long axis, and rotation angle were obtained. We verified the effectiveness of the proposed method by various simulation experiments and confirmed actual feasibility of the proposed method by applying to the real melon.

• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 2000.11c
• /
• pp.698-706
• /
• 2000

It is very difficult to mechanize tomato harvesting because identifying a tomato partly covered with leaves and stalks is not easy. This research was conducted to develop tomato harvesting robot which can identify a target tomato, determine its three dimensional position, and harvest it in a limited time. Followings were major findings in this study. The first visual system of the harvesting robot was composed of two CCD cameras, however, this could not detect tomatoes which are not seen on the view finder of the camera especially those partly covered by leaves or stalks. The second visual device, combined with two CCD cameras and pan/tilt procedures was designed to minimize the positioning errors within ${\pm}10mm$, but this is still not enough to detect tomatoes partly covered with leaves etc. Finally, laser distance detector was added to the visual system that could reduce the position detecting errors within 10mm in X-Y direction and 5mm in Z direction for the partly covered tomatoes.