Enhancing detection performance for robotic harvesting systems through RandAugment

Detecting crops accurately is the key challenge for harvesting robots, and deep learning methods are commonly used for this purpose. However, these methods require a large amount of training data, which can be a constraint. Data augmentation is frequently recommended to enlarge the size of training data and enhance the performance of the detection models. In this paper, we propose the use of RandAugment (RA) to improve crop detection performance by applying geometric, photometric, and partial occlusion transformations. We experimented with different transformation lists using popular detection networks on Tomato datasets. YOLOv3 with geometric transformations and partial occlusion achieved the highest accuracy of 76.42, showing a 3.47 increase over the baseline model without augmentation. RA also improved detection accuracy on other public datasets such as Apple, Kiwi, and Mango. We found that the optimal combination of transformations varied for each network to achieve the best performance. Additionally, we implemented YOLOv3 with the optimal transformation list to demonstrate the robotic tomato harvesting system and achieved a computing time of less than 60 ms from capturing the camera image to receiving the detection results.

Automated summary

Detecting crops accurately is a major challenge for harvesting robots, and deep learning methods are commonly used to address this issue. However, these methods require a large quantity of training data, which can be limiting. To improve crop detection performance, this study proposes using RandAugment (RA) to apply geometric, photometric, and partial occlusion transformations. The results show that YOLOv3 with RA achieved the highest accuracy of 76.42 on Tomato datasets, with a 3.47 increase compared to the baseline model without augmentation. RA also improved detection accuracy on other public datasets such as Apple, Kiwi, and Mango.