When uncrewed agricultural machinery performs autonomous operations in the field, it inevitably encounters obstacles such as persons, livestock, poles, and stones. Therefore, accurate recognition of obstacles in the field environment is an essential function. To ensure the safety and enhance the operational efficiency of autonomous farming equipment, this study proposes an improved YOLOv8-based field obstacle detection model, leveraging depth information obtained from binocular cameras for precise obstacle localization. The improved model incorporates the Large Separable Kernel Attention (LSKA) module to enhance the extraction of field obstacle features. Additionally, the use of a Poly Kernel Inception (PKI) Block reduces model size while improving obstacle detection across various scales. An auxiliary detection head is also added to improve accuracy. Combining the improved model with binocular cameras allows for the detection of obstacles and their three-dimensional coordinates. Experimental results demonstrate that the improved model achieves a mean average precision (mAP) of 91.8%, representing a 3.4% improvement over the original model, while reducing floating-point operations to 7.9 G (Giga). The improved model exhibits significant advantages compared to other algorithms. In localization accuracy tests, the maximum average error and relative error in the 2–10 m range for the distance between the camera and five types of obstacles were 0.16 m and 2.26%. These findings confirm that the designed model meets the requirements for obstacle detection and localization in field environments.

In China’s Yangtze River and Yellow River basin cotton-growing regions, the complex terrain, scattered planting areas, and poor adaptability of the existing machinery have led to a mechanized cotton harvesting rate of less than 10%. To address this issue, we designed a crawler chassis for a brush-roller cotton harvester. It is specifically tailored to meet the 76 cm row spacing agronomic requirement. We also conducted a theoretical analysis of the power transmission system for the crawler chassis. Initially, we considered the terrain characteristics of China’s inland cotton-growing regions and the current cotton agronomy practices. Based on these, we selected and designed the power system and chassis; then, a finite element static analysis was carried out on the chassis frame to ensure safety during operation; finally, field tests on the harvester’s operability, stability, and speed were carried out. The results show that the inverted trapezoidal crawler walking device, combined with a hydraulic continuously variable transmission and rear-drive design, enhances the crawler’s passability. The crawler parameters included a ground contact length of 1650 mm, a maximum ground clearance of 270 mm, a maximum operating speed of 6.1 km/h, and an actual turning radius of 2300 mm. The maximum deformation of the frame was 2.198 mm, the deformation of the walking chassis was 1.0716 mm, the maximum equivalent stress was 216.96 MPa, and the average equivalent stress of the entire frame was 5.6356 MPa, which complies with the physical properties of the selected material, Q235. The designed cotton harvester crawler chassis features stable straight-line and steering performance. The vehicle’s speed can be adjusted based on the complexity of the terrain, with timely steering responses, minimal compaction on cotton, and reduced soil damage, meeting the requirements for mechanized harvesting in China’s inland small plots.

Unmanned agricultural machinery can significantly reduce labor intensity while substantially enhancing operational efficiency and production benefits. However, the presence of various obstacles in complex farmland environments is inevitable. Accurate and efficient obstacle recognition technology, along with a reliable safety warning system, is a crucial prerequisite for ensuring the safe and stable operation of unmanned agricultural machinery. This study proposes a lightweight model for farmland obstacle detection by improving the YOLOv8n object detection algorithm. Specifically, we introduce the Context-Guided Block (CG Block) in the C2f module and the Context-Guide Fusion Module (CGFM) in the Feature Pyramid Network (FPN) to enhance the model’s contextual information perception during feature extraction and fusion. Additionally, we employ a Lightweight Shared Convolutional Separable Batch Normalization Detection Head in the detection head, which significantly reduces the number of parameters while improving detection accuracy. Experimental results demonstrate that our method achieves a mean average precision (mAP) of 92.3% at 59.1 frames per second (FPS). The improved model reduces parameter count and computational complexity by 31.9% and 33.4%, respectively, with a model size of only 4.2 MB. Compared to other algorithms, the proposed model maintains an optimal balance between parameter efficiency, computational cost, detection speed, and accuracy, exhibiting distinct advantages. Furthermore, we propose a safety warning strategy based on the relative velocity and distance between obstacles and the unmanned agricultural machinery. Field experiments conducted under this strategy reveal an overall warning accuracy of up to 86%, verifying the reliability of the safety warning system. This ensures that unmanned agricultural machinery can effectively mitigate potential safety risks during field operations.