This study proposes an improved algorithm based on the You Only Look Once v7 (YOLOv7) to address the low accuracy of apple fruit target recognition caused by high fruit density, occlusion, and overlapping issues. Firstly, we proposed a preprocessing algorithm for the split image with overlapping to improve the robotic intelligent picking recognition accuracy. Then, we divided the training, validation, and test sets. Secondly, the MobileOne module was introduced into the backbone network of YOLOv7 to achieve parametric fusion and reduce network computation. Afterward, we improved the SPPCSPS module and changed the serial channel to the parallel channel to enhance the speed of image feature fusion. We added an auxiliary detection head to the head structure. Finally, we conducted fruit target recognition based on model validation and tests. The results showed that the accuracy of the improved YOLOv7 algorithm increased by 6.9%. The recall rate increased by 10%, the mAP1 algorithm increased by 5%, and the mAP2 algorithm increased by 3.8%. The accuracy of the improved YOLOv7 algorithm was 3.5%, 14%, 9.1%, and 6.5% higher than that of other control YOLO algorithms, verifying that the improved YOLOv7 algorithm could significantly improve the fruit target recognition in high-density fruits.

Due to the short time, high labor intensity and high workload of fruit and vegetable harvesting, robotic harvesting instead of manual operations is the future. The accuracy of object detection and location is directly related to the picking efficiency, quality and speed of fruit-harvesting robots. Because of its low recognition accuracy, slow recognition speed and poor localization accuracy, the traditional algorithm cannot meet the requirements of automatic-harvesting robots. The increasingly evolving and powerful deep learning technology can effectively solve the above problems and has been widely used in the last few years. This work systematically summarizes and analyzes about 120 related literatures on the object detection and three-dimensional positioning algorithms of harvesting robots over the last 10 years, and reviews several significant methods. The difficulties and challenges faced by current fruit detection and localization algorithms are proposed from the aspects of the lack of large-scale high-quality datasets, the high complexity of the agricultural environment, etc. In response to the above challenges, corresponding solutions and future development trends are constructively proposed. Future research and technological development should first solve these current challenges using weakly supervised learning, efficient and lightweight model construction, multisensor fusion and so on.

Berries are popular by consumers for improving vision, lowering blood sugar, improving circulation, and cardiovascular protection. They are usually small, thin-skinned, and fragile, with inconsistent ripening times. Harvesting robots are able to accurately determine the ripeness of fruits, avoiding pulp breakage and nutrient loss caused by manual squeezing. This work reviews the development and application of berry harvesting robots with market prospects in recent years. Next, this paper discusses the key technologies of berry picking robots, including fruit detection and localization technology, motion planning technology, and end-effector and harvesting mechanism. It also discusses the challenges currently faced in the development of berry harvesting robots, including external factors such as unstructured working environments and internal technical difficulties such as robot design and control. To address these challenges, future berry picking robots should focus on developing weak supervision recognition models based on deep learning, high-speed collision-free multi-arm collaborative harvesting technology, and high fault-tolerant harvesting technology to improve picking efficiency and quality, reduce fruit damage, and promote the automation and intelligence of the berry harvesting.

As a vector of pine wood nematode disease, Monochamus alternatus is wood-boring, cryptic, and extremely difficult to control—features that have made them a disastrous problem. In this study, we present a method of scanning the galleries of Monochamus alternatus using CT (computed tomography) technology to obtain their systematic structure via 3D (three-dimensional) reconstruction, so as to clarify the gallery types and their structural parameters. TLC (thin-layer chromatography) scanning on wood segments damaged by M. alternatus was performed using a 128-row spiral CT GE Revolution EVO to obtain 64-layer CT scanned images. From the scanned images, we were able to clearly identify the beetle larvae and their galleries. The galleries were clearly delineated from the peripheral xylem, except for parts that were blocked by a frass–feces mixture, which were slightly blurred. Three-dimensional reconstruction of the galleries showed that most of the gallery types were C-shaped, and a few were S-shaped or Y-shaped. There was only one larva per gallery, and the galleries were separate. The vicinity of the entrance hole and the anterior part of the pupal chamber were blocked with a frass–feces mixture. There were no significant differences among the galleries’ parameters, such as the width of the entrance holes, tunneling depth, vertical length, blockage length and volume, total length of the galleries, and boring volume. With MIMICS (Materialise’s interactive medical image control system) image processing software, the images of each layer were made into a composite image, providing an effective way to visualize the 3D distribution of galleries. Using the methodology outlined in this study, both a single gallery structure and the spatial distribution of multiple galleries of M. alternatus can be shown, and the specific parameters of galleries can also be accurately calculated, which provides new ideas and methods for carrying out ecological and scientific research and precise prevention and control techniques of M. alternatus.

The characteristics of soil ditching and backfilling are crucial for orchard ditching operations. However, experimentally investigating the dynamic ditching and backfilling process is currently not feasible. To address this issue, the 3DGZ-50A self-propelled orchard ditching machine (SPODM) was designed using a modular concept, incorporating three types of ditching cutter discs (01#, 02#, and 03#). These discs were designed, trial-manufactured, and tested in orchard ditching experiments. A corresponding simulation model was also constructed using EDEM 2022 software. This study evaluated the ditching and backfilling process, analyzing the performance of the three cutter discs through experimental and simulation methods. Results indicated that the 01# and 02# cutter discs created V-shaped furrows, whereas the 03# cutter disc formed an arc-shaped furrow. The relative errors in the final furrow depth (D_f) and width (W_f) between experimental and simulated results were 30.70% and 8.61%, respectively, while those in the maximum furrow depth (D_m) and width (W_m) were 9.44% and 3.00%. These minor relative errors confirmed the accuracy of the simulation model. Regarding maximum power consumption, the 01# cutter disc used 86.3% of the power consumed by the 02# cutter disc and 85.1% of that used by the 03# cutter disc. During the ditching process, the blades penetrated the soil to create the maximum furrow cross-section, which then gradually decreased due to backfilling. Both simulation and test results demonstrated that the 01# cutter disc performed best, achieving a maximum furrow cross-sectional area (46.70%), minimum final surface furrow cross-sectional area (6.04%), and lower power consumption (31.03 kW). This study provides equipment for ditching operations in low-height close-planting orchards in northern China.

Intelligent apple-harvesting robots use a staggered distribution of branches and leaves during operation, causing problems such as slow motion planning, low operational efficiency, and high path cost for multi-degrees-of-freedom (DOF) harvesting manipulators. This study presents an autonomous apple-harvesting robotic arm-hand composite system that aims to improve the operational efficiency of intelligent harvesting in dwarf anvil-planted apple orchards. The machine vision system for fruit detection uses the deep learning convolutional neural network (CNN) YOLOv7 and RGB-D camera online detection coupling technology to rapidly recognise apples. The spatial depth information of the fruit area was then extracted from the aligned depth image for precise positioning. Coordinate transformation was used to obtain the coordinates of the fruit under the coordinate system of the manipulator. Based on the informed rapid-exploration random tree (Informed-RRT*) algorithm and path-planning model, the identified target apples were harvested without collision path planning. In an apple-harvesting test, the recognition accuracy of the visual system was 89.4%, and the average time to harvest a single apple was 9.69 s, which was 4.8% faster than the mainstream general harvesting technology. Moreover, the harvesting time for a single apple was reduced by 1.7%. Thus, the proposed system enabled accurate and efficient fruit harvesting.