In precision agriculture, after vegetable transplanters plant the seedlings, field management during the seedling stage is necessary to optimize the vegetable yield. Accurately identifying and extracting the centerlines of crop rows during the seedling stage is crucial for achieving the autonomous navigation of robots. However, the transplanted ridges often experience missing seedling rows. Additionally, due to the limited computational resources of field agricultural robots, a more lightweight navigation line fitting algorithm is required. To address these issues, this study focuses on mid-to-high ridges planted with double-row vegetables and develops a seedling band-based navigation line extraction model, a Seedling Navigation Convolutional Neural Network (SN-CNN). Firstly, we proposed the C2f_UIB module, which effectively reduces redundant computations by integrating Network Architecture Search (NAS) technologies, thus improving the model’s efficiency. Additionally, the model incorporates the Simplified Attention Mechanism (SimAM) in the neck section, enhancing the focus on hard-to-recognize samples. The experimental results demonstrate that the proposed SN-CNN model outperforms YOLOv5s, YOLOv7-tiny, YOLOv8n, and YOLOv8s in terms of the model parameters and accuracy. The SN-CNN model has a parameter count of only 2.37 M and achieves an mAP@0.5 of 94.6%. Compared to the baseline model, the parameter count is reduced by 28.4%, and the accuracy is improved by 2%. Finally, for practical deployment, the SN-CNN algorithm was implemented on the NVIDIA Jetson AGX Xavier, an embedded computing platform, to evaluate its real-time performance in navigation line fitting. We compared two fitting methods: Random Sample Consensus (RANSAC) and least squares (LS), using 100 images (50 test images and 50 field-collected images) to assess the accuracy and processing speed. The RANSAC method achieved a root mean square error (RMSE) of 5.7 pixels and a processing time of 25 milliseconds per image, demonstrating a superior fitting accuracy, while meeting the real-time requirements for navigation line detection. This performance highlights the potential of the SN-CNN model as an effective solution for autonomous navigation in field cross-ridge walking robots.

The rapid and accurate detection of broccoli seedling planting quality is crucial for the implementation of robotic intelligent field management. However, existing algorithms often face issues of false detections and missed detections when identifying the categories of broccoli planting quality. For instance, the similarity between the features of broccoli root balls and soil, along with the potential for being obscured by leaves, leads to false detections of “exposed seedlings”. Additionally, features left by the end effector resemble the background, making the detection of the “missed hills” category challenging. Moreover, existing algorithms require substantial computational resources and memory. To address these challenges, we developed Seedling-YOLO, a deep-learning model dedicated to the visual detection of broccoli planting quality. Initially, we designed a new module, the Efficient Layer Aggregation Networks-Pconv (ELAN_P), utilizing partial convolution (Pconv). This module serves as the backbone feature extraction network, effectively reducing redundant calculations. Furthermore, the model incorporates the Content-aware ReAssembly of Features (CARAFE) and Coordinate Attention (CA), enhancing its focus on the long-range spatial information of challenging-to-detect samples. Experimental results demonstrate that our Seedling-YOLO model outperforms YOLOv4-tiny, YOLOv5s, YOLOv7-tiny, and YOLOv7 in terms of speed and precision, particularly in detecting ‘exposed seedlings’ and ‘missed hills’-key categories impacting yield, with Average Precision (AP) values of 94.2% and 92.2%, respectively. The model achieved a mean Average Precision of 0.5 (mAP@0.5) of 94.3% and a frame rate of 29.7 frames per second (FPS). In field tests conducted with double-row vegetable ridges at a plant spacing of 0.4 m and robot speed of 0.6 m/s, Seedling-YOLO exhibited optimal efficiency and precision. It achieved an actual detection precision of 93% and a detection efficiency of 180 plants/min, meeting the requirements for real-time and precise detection. This model can be deployed on seedling replenishment robots, providing a visual solution for robots, thereby enhancing vegetable yield.

The operational performance of cereal seeding machinery influences the yield and quality of cereals. In this article, we review the existing literature on intelligent technologies for cereal seeding machinery, encompassing active controllable seeding actuators, intelligent seeding rate control, and intelligent seed position control systems. In this manuscript, (1) the characteristics and innovative structures of existing motor-driven seed-metering devices and ground surface profiling mechanisms are expounded; (2) state-of-the-art detection principles and applications for soil property sensors are described based on different soil properties; (3) optimal seeding rate decision approaches based on soil properties are summarized; (4) the research state of seeding rate measuring and control technologies is expounded in detail; (5) trajectory control methods for seeding machinery and seeding depth control systems are described based on measurement and control principles; and (6) the present state, limitations, and future development directions of intelligent cereal seeding machinery are described. In the future, more advanced multi-algorithm and multi-sensor fusion technologies for soil property detection, optimal seeding rate decisions, seeding rates, and seed position control are likely to evolve. This review not only expounds the latest studies on intelligent actuating, sensing, and control technologies for intelligent cereal seeding machinery, but also discusses the shortcomings of existing intelligent seeding technologies and future developing trends in detail. This review, therefore, offers a reference for future research in the domain of intelligent seeding machinery for cereals.

To address the inefficiency and instability of automatic transplanting machines, a dual-row seedling pick-up device and its corresponding control system were developed. Existing seedling end-effectors are primarily mechanically controlled, and the seedling needles can easily cause damage to the interior of the bowl. In order to reduce the damage inflicted by the end-effectors to the bowl, this paper conducted a mechanical analysis of the end-effector. At the same time, a buffer optimization analysis was conducted on the operation of the end-effector, and a flexible pneumatic end-effector for seedling picking was designed. The control system combined the detection of multiple sensors to monitor the process of seedling picking and throwing. By coordinating the lifting cylinder and clamping cylinder, the system effectively reduced seedling pot damage while improving seedling picking efficiency. By setting the operating parameters of the servo motor, the goal of low-speed and high-efficiency seedling picking was achieved. To evaluate the performance of the control system, the linear displacement sensors and acceleration testing systems were used to analyze the performance of the seedling throwing. The results showed that the seedling picking efficiency could reach 180 plants min⁻¹, with no significant difference between the actual measured moving distance and the theoretical setting distance. The positioning error remained stable between 0.5 and 0.9 mm, which met the requirements for seedling picking accuracy. The buffer optimization design reduced the peak acceleration of the end-effector from −22.1 m/s² to −13.4 m/s², and the peak value was reduced by 39.4%, which proved the significant effectiveness of the buffer design. A performance test was conducted using 128-hole seed trays and 33-day-old cabbage seedlings for seedling picking and throwing. When the planting frequency reached 90 plants/row·min⁻¹, the average success rate of picking and throwing seedlings was 97.3%. This indicates that the various components of the designed seedling pick-up device work in good coordination during operation, and the control system operates stably. Technical requirements for the automatic mechanical transplanting of tray seedlings were achieved, which can provide reference for research on automatic transplanting machines.

There are various types of fruits and vegetables that need to be planted on ridges. In order to allow for seedlings with a certain row space and seedling space, the ridge transplanter should be able to track along the ridge. Therefore, an ultrasonic ridge-tracking method and system were developed to let the ridge transplanter track the ridge accurately. The ultrasonic ridge-tracking method mainly contains a limiter sliding window filtering algorithm and a fuzzy look-ahead distance decision model. The limiter sliding window filtering algorithm was proposed to filter the abnormal measuring results to avoid disoperation of the steering mechanism. Moreover, the fuzzy look-ahead distance decision model was proposed to determine the optimal look-ahead distance in order to obtain a desirable tracking performance. Additionally, a comparison experiment of the proposed ultrasonic ridge-tracking method and the universal pure pursuit method was conducted. The experimental results show that the greatest mean absolute errors of the lateral deviations of the ultrasonic ridge-tracking method and universal pure pursuit were 10.56 mm and 13.11 mm. The greatest maximum absolute errors of the lateral deviations of the ultrasonic ridge-tracking method and universal pure pursuit were 18.87 mm and 23.23 mm. In addition, the greatest root mean square error of the lateral deviation of the ultrasonic ridge-tracking method and the universal pure pursuit method were 13.52 mm and 15.66 mm. According to the ridge-tracking performance of the proposed ultrasonic ridge-tracking method, it can be used in practical transplanting conditions. Moreover, in other fields, robots or intelligent machinery can also apply the proposed ultrasonic ridge-tracking method to track objects similar to ridges.

This paper aims to address the problem of the low success rate of seedling picking and throwing, and the high damage rate of pot seedling, caused by the unclear interaction and parameter mismatch between the seedling picking end effector and the pot seedling during the seedling picking and throwing process of automatic transplanters. An EDEM–RecurDyn coupled simulation was conducted, through which the disturbance of substrate particles in the bowl body during the inserting, extracting, and transporting processes by the seedling picking end effector was visualized and analyzed. The force and motion responses of the particles during their interaction with the seedling picking end effector were explored, and the working parameters of the seedling picking end effector were optimized. A seedling picking end effector using two fingers and four needles is taken as the research object, a kinematic mathematical model of the seedling picking end effector is established, and the dimensional parameters of each component of the end effector are determined. Physical characteristic tests are conducted on Shanghai bok choy pot seedlings to obtain relevant parameters. A discrete element model of the pot seedling is established in EDEM 2022 software, and a virtual prototype model of the seedling picking end effector is established in Recurdyn 2024 software. Through EDEM-Recurdyn coupled simulation, the force and movement of the substrate particles in the bowl body during the inserting, extracting, and transporting processes of the seedling picking end effector under different operating parameters were explored, providing a theoretical basis for optimizing the working parameters of the end effector. The inserting and extracting velocity, transporting velocity, and inserting depth of the seedling picking end effector were used as experimental factors, and the success rate of seedling picking and throwing, and the loss rate of substrate, were used as evaluation indicators; single-factor tests and three-factor, three-level Box–Behnken bench tests were conducted. Variance analysis, response surface methodology, and multi-objective optimization were performed using Design-Expert 13 software to obtain the optimal parameter combination: when the inserting and extracting velocity was 228 mm/s, the transporting velocity was 264 mm/s, the inserting depth was 37 mm, the success rate of seedling picking and throwing was 97.48%, and the loss rate of substrate was 2.12%. A verification experiment was conducted on the bench, and the success rate of seedling picking and throwing was 97.35%, and the loss rate of substrate was 2.34%, which was largely consistent with the optimized results, thereby confirming the rationality of the established model and optimized parameters. Field trial showed the success rate of seedling picking and throwing was 97.04%, and the loss rate of substrate was 2.41%. The error between the success rate of seedling picking and throwing and the optimized result was 0.45%, indicating that the seedling picking end effector has strong anti-interference ability, and verifying the feasibility and practicality of the established model and optimized parameters.