to perform inverse kinematics calculations on the joints and angles ( see Figure 4A ). Ultimately , the COBOT successfully completes the harvesting process by manipulating the end effector , a flexible gripper , to the specified tomato location based on the inverse kinematics calculation data for each joint rotation angle , and then proceeds to harvest it ( see Figure 4B ). The entire process can be viewed and altered remotely by wireless connection , such as WIFI and internet , using a computer , tablet , or smartphone . The probability of successfully reaching the tomato exceeds 80 %. The performance of machine vision systems in recognising and locating fruit can be influenced by the unpredictable and changeable illumination conditions in the field environment , as well as the different and complex canopy structures [ 18 ] and the changing colour , shape , and size of the fruit . Moreover , the capacity to precisely identify fruit in canopy photographs is greatly impeded by the hindrance provided by leaves , branches , and other fruits [ 11 ]. In the near future , the success rate could be enhanced by incorporating AI through the process of machine learning . This would involve training the system to overcome variations in lighting conditions and enabling it to identify objects even when they are partially obstructed .
Fruit Grading Robot
MARDI has created a robotic arm system ( see Figure 5 ) specifically designed to perform the task of grading watermelons depending on their weight . The robotic arm system comprises a robotic arm , automated control software , an electronic “ load cell ” scale , and a table designed for watermelon collection depending on weight . The robotic arm system is equipped with a “ suction cup ” that is utilised to elevate and transport watermelons from the digital weighing area ( see Figure 5A ) to the grading table ( see Figure 5B ). The utilisation of this robotic arm technology becomes advantageous in the efficient manipulation of weighty watermelons , particularly in the laborious task of manually grading post-harvest watermelon produce . This method minimises the likelihood of physical harm to watermelons as compared to manual handling . This robotic arm technology can also decrease the spatial need of the machine floor area when
A
Figure 5 : Watermelon grading robotic system : ( A ) The robotic arm system lifts the watermelon on the grading table , ( B ) The robotic arm system transfers watermelon to the pre-determined weight slot
compared with using a “ conveyor ” system for transferring watermelons from the weighing section to the watermelon collection section .
The robotic arm system follows the Standard Fresh Water Melon MS 1028 ( 2005 ) for evaluating watermelons , which is based on a weight size grading scheme . The practice of classifying watermelons by weight has been devised to fulfil the export standards for watermelons destined for international markets , particularly Dubai , Hong Kong , and European countries . The watermelons will be sorted and packaged into boxes based on the specific size requirements set by the export market .
Employing robotic arm technology for watermelon grading can address the issues of labour scarcity and reliance on foreign workers . It can assist in managing the weighty watermelon , and the technique does not have any adverse effects on the quality of the watermelons . This technology can also be utilised by entrepreneurs with restricted workspaces . The robotic arm system for watermelon grading is more space-
B
47