Corn is one of the most important crops for national food security. However, weed is a major influencing factor of biological stress on corn growth. They compete with corn seedlings for water, nutrients, and light, serve as hosts for pests and pathogens, thereby causing declines in both crop yield and quality. Conventional weed control strategies rely typically on manual identification or extensive herbicide application, both of which are associated with low efficiency, resource waste, and environmental pollution. Existing detection models for corn seedlings and weeds often suffer from large parameter counts, high computational costs, and low accuracy in complex agricultural environments, leading to the false positives and missed detections. In this study, a lightweight detection framework named YOLO11-SAW was developed using YOLO11n. The high accuracy and inference efficiency were achieved well-suitable for deployment on edge devices. Specifically, three aspects were proposed for the improved YOLO11n: Backbone feature extraction, neck feature fusion, and the loss function for bounding box regression. 1) C3STR module was constructed to combine the C3 module with the Swin Transformer at the end of the backbone network. Global contextual dependencies were better captured to distinguish corn seedlings from weeds in scenarios with the overlapping leaves, target occlusion, dense distribution, and complex soil backgrounds. 2) Neck structure was augmented with the Alterable Kernel Convolution (AKConv) module. The learnable offsets were introduced to replace conventional convolution. The adaptability to multi-scale and deformed weed targets were enhanced after modification to simultaneously reduce both floating-point operations (FLOPs) and model parameters. 3) A bounding box regression was refined to introduce the WIoUv3 loss function. Training instability caused by fixed gradient magnitudes was alleviated to enhance the suppression of low-quality samples. The high-quality samples were consequently promoted the convergence stability and localization accuracy. A series of comparative experiments were conducted to evaluate the effectiveness of the improved model on a self-constructed corn seedling and weed dataset. The YOLO11-SAW was compared with several representative models under unified experiment, including Faster R-CNN, Swin Transformer, RT-DETR-L, and multiple YOLO-series models. It was found that the YOLO11-SAW outperformed existing models in both detection accuracy and computational efficiency. Compared with the baseline YOLO11n model, YOLO11-SAW improved Precision by 1.71 percentage points and Recall by 2.18 percentage points, with mAP_0.5 and mAP_0.5:0.95 reaching 96.40% and 76.17%, respectively, indicating strong detection and localization. In terms of model complexity, the parameters, floating-point operations and model size of YOLO11-SAW are 1.97 M, 8.0 G and 4.3 MB respectively. In addition, an inference speed of 202.77 frames per second (FPS) was fully met the real-time requirements of intelligent agricultural applications. In summary, the YOLO11-SAW model can real-time and accurately detect the corn seedlings and weeds in complex agricultural environments. The findings can provide practical technical support to deploy the mobile robots and variable-rate spraying on edge devices in smart agriculture.

Obstacle detection and avoidance are critical technologies in the autonomous navigation of agricultural robots. The ability to accurately perceive and avoid obstacles ensures the safe operation of agricultural robots in complex environments, reduces labor costs, and improves production efficiency. This paper provides a comprehensive review of the applications of single-sensor technologies and multi-sensor fusion technologies in obstacle detection for agricultural robots from various perspectives. First, the paper reviews recent advancements in using single-sensor technologies for obstacle detection, including RGB cameras, panoramic cameras, stereo cameras, depth cameras, LiDAR, ultrasonic radar, and millimeter-wave radar. The technical advantages and limitations of each type of sensor are analyzed in detail. For example, RGB and depth cameras offer high spatial resolution and are effective in structured environments but face challenges in low-light or high-dust conditions. Similarly, LiDAR provides precise distance measurements and is highly reliable for obstacle detection but remains costly for large-scale deployment. Ultrasonic and millimeter-wave radars, known for their robustness in harsh conditions, are often limited by their resolution and detection range. These analyses highlight the trade-offs involved in adopting specific sensors for agricultural robotics. Second, the paper summarizes the research progress in multi-sensor fusion technologies for obstacle detection, such as the integration of visual sensors with LiDAR and the fusion of visual sensors with millimeter-wave radar. Detailed analyses are provided on the characteristics and advantages of these fusion techniques. For instance, combining vision-based sensors with LiDAR enhances obstacle detection accuracy by leveraging the complementary strengths of high-resolution imagery and precise distance measurements. Similarly, the fusion of millimeter-wave radar with vision-based systems enables reliable detection in adverse weather conditions, such as rain or fog, while mitigating the limitations of individual sensors. These fusion approaches are discussed in the context of their suitability for various agricultural scenarios, emphasizing their ability to enhance the robustness and reliability of obstacle detection systems. Additionally, the paper reviews the progress of obstacle avoidance technologies for agricultural robots. Agricultural robots are broadly categorized into three types: large-scale agricultural machinery, small agricultural robots, and agricultural drones. For each category, the paper systematically examines the recent advancements in obstacle avoidance technologies, focusing on their applications in complex agricultural environments. The discussion highlights the unique requirements and challenges faced by each type of robot. For instance, large-scale agricultural machinery requires highly reliable obstacle avoidance systems to navigate expansive fields efficiently, while small robots emphasize flexibility and precision in row crops. Agricultural unmanned aerial vehicle, operating in three-dimensional environments, face unique challenges in real-time obstacle detection and avoidance due to their high-speed motion and variable terrain. Finally, the paper summarizes the current state of research and identifies the key challenges in obstacle detection and avoidance technologies for agricultural robots. These include the need for improving the adaptability of sensors and algorithms to highly dynamic agricultural environments, reducing the cost and energy consumption of sensing systems, and enhancing the robustness of multi-sensor fusion approaches. Furthermore, this paper provides a forward-looking perspective on the development of these technologies, offering theoretical foundations and technical references to accelerate the advancement of autonomous navigation for agricultural robots. This comprehensive review aims to contribute to the field by identifying current limitations and exploring innovative solutions, ultimately facilitating the safe and efficient deployment of agricultural robots in diverse and challenging environments.