Search

Vegetables have the largest planting areas besides grains. Machine vision navigation has been one of the most crucial indicators of mechanization, automation, and intelligence in modern agriculture. Most vegetable transplanters are still manually driven at present. It is also necessary to detect the ridge mounds before the navigation of the transplanter. Since the ridge mounds are often free of crops before transplanting, it is still challenging to find references. Therefore, it is a high demand to extract the navigation lines with crop-free ridges mounds under complex scenes. There was also similar color information and small texture difference in crop-free ridges rows. Traditional image processing cannot fully meet large-scale production. In this study, a ridge row segmentation model was proposed using an improved version of DeepLabV3+. The real-time performance of semantic segmentation was also achieved with the high applicability, accuracy, and detection speed. The traditional DeepLabV3+ network was simplified to replace the Xception backbone network with the MobileNetV2 network. The speed of detection and the real-time performance were obtained after that. The DeepLabV3+ model incorporated the Convolutional Block Attention Module attention mechanism, in order to better treat the ridge boundary information. The important details of the ridge boundary were focused to accurately detect and classify the target objects. Navigational feature points were obtained using the ridge boundary information. In cases where the seedlingless ridges were present, the navigational feature points were deviated from the intended positions. Accordingly, the feature points were adjusted for the guidance of accurate navigation. The quartiles were utilized to filter out any outliers among the navigation feature points. Any data points were identified and removed to deviate significantly from the norm. In addition, the least squares method was used to fit the navigation line using the fitted feature points. A reliable reference of the navigation line was then obtained to compensate for any deviations from the seedlingless ridges. Overall, the simplified DeepLabV3+ network with the MobileNetV2 backbone was incorporated with the CBAM attention mechanism. There were the high detection speed, real-time performance and accurate navigations, even in the challenging scenarios with the ridge boundaries. Two locations were also selected from the images, in order to improve the applicability of the model in the environments of crop-free ridges. The challenge remained on the different soil qualities, lighting conditions and seedlingless ridges in the field test. The dataset consisted of 1350 images in the training set, and 150 images in the validation set. The images were then expanded using data enhancement. The results indicate that the improved model was achieved with an average pixel accuracy of 96.27%, an average intersection and merger ratio of 93.18%, and an average detection frame rate of 84.21 frames per second. The average intersection and merger ratio, the average pixel accuracy, and the frame rate were improved by 1.78 percentage points, 0.83 percentage points, and 29.32, respectively, compared with the pre-improvement period. The best performance was then achieved in the MobileNetV2, thus meeting the requirements for navigation recognition, compared with the various backbone networks. On average, the pixel accuracies of the model were improved by 0.83 and 3.28 percentage points, respectively, compared with the Xception and MobileNetV3. Furthermore, the improved model also demonstrated better average accuracy, average intersection ratio, and frame rate than PSPNet, U-Net, HRNet, Segformer, and DeepLabV3+. The Hough transform and Random Sample Consensus were much less effective in obtaining the navigation lines from different scenes, compared with the maximum angular error of 1.2° and the maximum pixel distance error of 9 in various ridges environments. These findings can serve as a strong reference for crop-free ridge navigation in agricultural robots, thus promoting the development of intelligent agricultural equipment. [ABSTRACT FROM AUTHOR]

An end-effector is one of the most important components in the fruit picking robot. The picking mechanism of the end-effector can also dominate the picking efficiency. However, the picking end-effector with the fixed structure and parameters cannot fully meet the picking needs of different fruit diameter ranges, due to the wide variety of objects being picked at present. In this study, an underactuated articulated end-effector was designed, including four picking fingers. Each picking finger with two joints was driven by a connecting rod to simulate the human envelope-picking action. The hand with the intention of picking citrus was used to simulate a similar spherical fruit. The better performance of the end-effector was achieved in the small number of system drives, the low overall system complexity, and the small damage to the fruit during the picking process. A parametric design was also proposed using an improved genetic algorithm (GA), in order to improve the repetitive design of this type of end-effector. The picking range of fruit diameters was expanded to improve the feasibility of the movement. Microsoft Visual Studio 2012 development platform was utilized to customize the MFC design parameter interface. The end-effector structure was then previewed to input the initial parameters in the interface. The improved GA optimization was adopted to combine with the secondary development module NXOpen of NX. The parameters were assigned to the mechanism model in the form of expressions during optimization. Specifically, the structural parameters were optimized during this time. The end-effector was pre-modeled to fully meet the optimal picking conditions. Adams platform was employed to conduct the dynamic simulation analysis on the optimized grasping end-effector model. The maximum displacement offset of the optimized end-effector increased by about 29.1%. The optimal displacement value was achieved in the objective function. The simulation curve image was added into the parametric design system for display through the Python script. The correctness of the optimal design was then verified after simulation. A physical prototype was carried out indoor and orchard citrus picking experiments, according to the designed end-effector mechanism. In the laboratory environment, the success rate of envelope picking reached 100% for 20 citruses with a diameter in the range of 64-102 mm. The end-effector was connected to the self-made robotic arm. The end of the mechanical arm turned the end-effector "wrist" after the finger mechanism formed an envelope, and then twisted the fruit handle to complete the picking. Picking experiments were performed on the citrus with the different fruit diameters in the orchard. The test results show that the success rate of picking citrus with a diameter of 68-106 mm reached 92.9%. The displacement parameters of the optimized end-effector were measured to obtain the maximum displacement (61.4 mm) of the single-ended joint, similar to the Adams simulation. Since the maximum offset was set in the optimization objective function, there was a threshold for the maximum opening of the picking node, in order to reduce the interference of branches and leaves. The total success rate reached 92.9% in the picking test. The parametric design can realize the rapid deformation response of the end-effector to the spherical fruits with different fruit diameter ranges. GA was also used to optimize the working parameters of the mechanism, in order to improve the feasibility of picking and the accuracy of mechanism design. [ABSTRACT FROM AUTHOR]

Inferior bowl seedlings have posed a great threat to the survival rate of later seedling transplantation in greenhouse hole tray seedling cultivation. It is urgent to remove the matrix blocks of these holes before leaving the factory. However, the existing mechanical removal often causes the particle scattering and omission. Air suction removal can be expected to effectively compensate for this defect. In this study, a series of experiments were conducted to optimize and then verify the structural parameters in the matrix air suction removal scheme. A systematic analysis was implemented to determine the relationship between pipeline design and adsorption efficiency under negative pressure. A single factor experiment was also carried out to preliminarily select the optimal values of pipe diameter (10, 18, 26, and 32 mm), pipe spacing (0, 4, 8, and 12 mm), and pipe length (1 000, 2 000, 3 000, and 4000 mm). A optimal combination was obtained with the pipe diameter 26 mm, pipe spacing 6 mm, and pipe length 2 000 mm. Further expansion was carried out with 200-hole potted seedlings as the research object (The seedling emergence rate of the entire tray was 92%, and the health rate was 86%). Four factors were performed on the substrate removal rate, including pipe type (round tube, flat tube, and square tube), pipe diameter (inner diameter of 23, 26, and 29 mm), pipe spacing (2, 6, and 10 mm), and pipe length (1 000, 2 000, and 3 000 mm). Response surface analysis was conducted as well. The experimental results show that all four factors posed an impact on the matrix removel rate. Firstly, the pipe spacing and the pip type shared the greatest impact on the matrix removel rate, and then the pipe diameter. Nevertheless, there was the little effect of pipe length on the cleaning. Reasonable scaling was carried out in the later experimental design. Secondly, the larger the pipe diameter was, the better performance was. It was also necessary to consider the limitation of hole space and the damage to adjacent hole healthy seedlings. The smaller the pipe spacing was, the greater the negative pressure capacity was, and the better the matrix removel rate was, but the suction cup phenomenon was more serious. There was the little impact of the pipe length on the overall suction, but it cannot be ignored. Finally, the optimal combination of various factors was achieved: a square tube, an inner diameter of 26 mm at the end of the straw, a distance of 3 mm from the end of the straw to the port on the hole, and a length of approximately 1000 mm. Consequently, 95% matrix removal was achieved at the maximum negative pressure of 22 kPa,which is an average increase of 2.94% compared to before optimization. More efficient substrate removal was achieved under the same negative pressure. The better performance was gained than the mechanical removal equipment of low-quality bowl seedling substrate. This finding and also provide the data reference for the design and development of removal equipment with substrate negative pressure. [ABSTRACT FROM AUTHOR]