A grafting robot needs to obtain the position information for the plant seedlings to perform automatic grafting operations. Accurately measuring the cutting points required during grafting plays a pivotal role in completing high-quality grafting tasks. Traditional visual detection models exhibit suboptimal performance on edge devices due to their large model size and suffer from limited detection efficiency. To achieve rapid and precise cutting point localization, this study proposes an all-new module termed the Stimulative Upsample Block (SUB). Additionally, the Spatial and Channel Reconstruction Convolution (SCConv) and a Local Importance-based Attention (LIA) mechanism are incorporated into the YOLO11n architecture, culminating in an enhanced model named YOLO11n-LSS. Our model achieved mean average precision (mAP) values of 93.2% for the instance segmentation task and 98.9% for the key point detection task. Compared to YOLOv8n and YOLO11n, our model reduces the number of parameters and computational cost by 4.6% and 3.8%, respectively, making it a high-performance and lightweight solution. The successful application of the new algorithm will significantly improve the production efficiency of automated tomato grafting and contribute to the advancement of the tomato cultivation industry.

Automatic guidance of agricultural vehicles requires automatic execution of operation commands received from the navigation controller by using electronically controlled mechanisms for wheel steering, speed changing and work implementing. Automatic steering contributes as a prerequisite technique in automatic and semi-automatic agricultural navigation. This research aimed to develop an electric automatic steering system that was compact in its structure and integrated into original steering mechanism in a simply and convenient way for aftermarket modification. A brushless motor and reducer assembly was utilized to provide an adequate steering torque instead of manual maneuver. A rapid assembling approach was proposed by passing the steering shaft through the hollow output shaft. A digital proportional-integral-differential (PID) algorithm was implemented to calculate the rotation speeds and directions by comparing the desired angle and the actual angle, which was implemented in a printed circuit board with a microcontroller unit (MCU) and interface chips. An unmanned wheeled tractor was applied as test platform to integrate the newly developed electric automatic steering system. Tests were conducted to evaluate its performance in terms of stability and responsiveness. An autonomous navigation system guided the tractor along target paths in the field by sending steering commands to the electric automatic steering system. The results show that the steering angle error was less than 0.81° when desired steering angle was less than 10°. The lateral error difference was no more than 4.76 cm when repeating following the same target path, which indicated that the electric automatic steering system responded accurately and robustly to steering commands.