Crop row detection technology, as one of the key technologies for agricultural robots to achieve autonomous navigation and precise operations, is related to the precision and stability of agricultural machinery operations. Its research and development will also significantly determine the development process of intelligent agriculture. The paper first summarizes the mainstream technical methods, performance evaluation systems, and adaptability analysis of typical agricultural scenes for crop row detection. The paper also summarizes and explains the technical principles and characteristics of traditional methods based on visual sensors, point cloud preprocessing based on LiDAR, line structure extraction and 3D feature calculation methods, and multi-sensor fusion methods. Secondly, a review was conducted on performance evaluation criteria such as accuracy, efficiency, robustness, and practicality, analyzing and comparing the applicability of different methods in typical scenarios such as open fields, facility agriculture, orchards, and special terrains. Based on the multidimensional analysis above, it is concluded that a single technology has specific environmental adaptability limitations. Multi-sensor fusion can help improve robustness in complex scenarios, and the fusion advantage will gradually increase with the increase in the number of sensors. Suggestions on the development of agricultural robot navigation technology are made based on the current status of technological applications in the past five years and the needs for future development. This review systematically summarizes crop row detection technology, providing a clear technical framework and scenario adaptation reference for research in this field, and striving to promote the development of precision and efficiency in agricultural production.

The resistance of ordinary potato digging shovels can increase dramatically when used in a clay soil because of the adhesion between the soil and shovel. In this paper, a new type of bionic potato digging shovel was designed to decrease adhesion. The bionic structural elements, i.e., scalelike units (S-U) were applied to the potato digging shovel with inspiration from pangolin scales. The discrete element method (DEM) considered cohesion was used to simulate the drag reduction performance in clayey soil conditions. An ordinary plane shovel (O-P-S) was used for comparison. Three indicators (total force, draft force and compressive force) were used to characterize the drag reduction performance. The effect of the design variables of the bionic structures (length [l] and height [h]) and the transversal and longitudinal arrangement spacing (S1 and S2) of the structures on the drag reduction performance were analyzed. The results showed that the drag reduction performance of the bionic shovels with suitable parameters was better than that of the O-P-S. The best bionic sample labeled as a bionic prototype had a 22.26% drag reduction rate during the soil bin test and a 14.19% drag reduction rate during the field test compared to the O-P-S.

Rose tea is a type of flower tea in China’s reprocessed tea category, which is divided into seven grades, including super flower, primary flower, flower bud, flower heart, yellow flower, scattered flower, and waste flower. Grading rose tea into distinct quality levels is a practice that is essential to boosting their competitive advantage. Manual grading is inefficient. We provide a lightweight model to advance rose tea grading automation. Firstly, four kinds of attention mechanisms were introduced into the backbone and compared. According to the experimental results, the Convolutional Block Attention Module (CBAM) was chosen in the end due to its ultimate capacity to enhance the overall detection performance of the model. Second, the lightweight module C2fGhost was utilized to change the original C2f module in the neck to lighten the network while maintaining detection performance. Finally, we used the SIoU loss in place of the CIoU loss to improve the boundary regression performance of the model. The results showed that the mAP, precision (P), recall (R), FPS, GFLOPs, and Params values of the proposed model were 86.16%, 89.77%, 83.01%, 166.58, 7.978, and 2.746 M, respectively. Compared with the original model, the mAP, P, and R values increased by 0.67%, 0.73%, and 0.64%, the GFLOPs and Params decreased by 0.88 and 0.411 M, respectively, and the speed was comparable. The model proposed in this study also performed better than other advanced detection models. It provides theoretical research and technical support for the intelligent grading of roses.