Due to their rapid development and wide application in modern agriculture, robots, mobile terminals, and intelligent devices have become vital technologies and fundamental research topics for the development of intelligent and precision agriculture. Accurate and efficient target detection technology is required for mobile inspection terminals, picking robots, and intelligent sorting equipment in tomato production and management in plant factories. However, due to the limitations of computer power, storage capacity, and the complexity of the plant factory (PF) environment, the precision of small-target detection for tomatoes in real-world applications is inadequate. Therefore, we propose an improved Small MobileNet YOLOv5 (SM-YOLOv5) detection algorithm and model based on YOLOv5 for target detection by tomato-picking robots in plant factories. Firstly, MobileNetV3-Large was used as the backbone network to make the model structure lightweight and improve its running performance. Secondly, a small-target detection layer was added to improve the accuracy of small-target detection for tomatoes. The constructed PF tomato dataset was used for training. Compared with the YOLOv5 baseline model, the mAP of the improved SM-YOLOv5 model was increased by 1.4%, reaching 98.8%. The model size was only 6.33 MB, which was 42.48% that of YOLOv5, and it required only 7.6 GFLOPs, which was half that required by YOLOv5. The experiment showed that the improved SM-YOLOv5 model had a precision of 97.8% and a recall rate of 96.7%. The model is lightweight and has excellent detection performance, and so it can meet the real-time detection requirements of tomato-picking robots in plant factories.

In order to realize the intelligent online yield estimation of tomato in the plant factory with artificial lighting (PFAL), a recognition method of tomato red fruit and green fruit based on improved yolov3 deep learning model was proposed to count and estimate tomato fruit yield under natural growth state. According to the planting environment and facility conditions of tomato plants, a computer vision system for fruit counting and yield estimation was designed and the new position loss function was based on the generalized intersection over union (GIoU), which improved the traditional YOLO algorithm loss function. Meanwhile, the scale invariant feature could promote the description precision of the different shapes of fruits. Based on the construction and labeling of the sample image data, the K-means clustering algorithm was used to obtain nine prior boxes of different specifications which were assigned according to the hierarchical level of the feature map. The experimental results of model training and evaluation showed that the mean average precision (mAP) of the improved detection model reached 99.3%, which was 2.7% higher than that of the traditional YOLOv3 model, and the processing time for a single image declined to 15 ms. Moreover, the improved YOLOv3 model had better identification effects for dense and shaded fruits. The research results can provide yield estimation methods and technical support for the research and development of intelligent control system for planting fruits and vegetables in plant factories, greenhouses and fields.

No abstract