Recognition and 3D positional estimation of apples during harvesting from a robotic platform in a moving vehicle are still challenging. Fruit clusters, branches, foliage, low resolution, and different illuminations are unavoidable and cause errors in different environmental conditions. Therefore, this research aimed to develop a recognition system based on training datasets from an augmented, complex apple orchard. The recognition system was evaluated using deep learning algorithms established from a convolutional neural network (CNN). The dynamic accuracy of the modern artificial neural networks involving 3D coordinates for deploying robotic arms at different forward-moving speeds from an experimental vehicle was investigated to compare the recognition and tracking localization accuracy. In this study, a Realsense D455 RGB-D camera was selected to acquire 3D coordinates of each detected and counted apple attached to artificial trees placed in the field to propose a specially designed structure for ease of robotic harvesting. A 3D camera, YOLO (You Only Look Once), YOLOv4, YOLOv5, YOLOv7, and EfficienDet state-of-the-art models were utilized for object detection. The Deep SORT algorithm was employed for tracking and counting detected apples using perpendicular, 15°, and 30° orientations. The 3D coordinates were obtained for each tracked apple when the on-board camera in the vehicle passed the reference line and was set in the middle of the image frame. To optimize harvesting at three different speeds (0.052 ms−1, 0.069 ms−1, and 0.098 ms−1), the accuracy of 3D coordinates was compared for three forward-moving speeds and three camera angles (15°, 30°, and 90°). The mean average precision (mAP@0.5) values of YOLOv4, YOLOv5, YOLOv7, and EfficientDet were 0.84, 0.86, 0.905, and 0.775, respectively. The lowest root mean square error (RMSE) was 1.54 cm for the apples detected by EfficientDet at a 15° orientation and a speed of 0.098 ms−1. In terms of counting apples, YOLOv5 and YOLOv7 showed a higher number of detections in outdoor dynamic conditions, achieving a counting accuracy of 86.6%. We concluded that the EfficientDet deep learning algorithm at a 15° orientation in 3D coordinates can be employed for further robotic arm development while harvesting apples in a specially designed orchard.

Poultry production utilizes many available technologies in terms of farm-industry automation and sanitary control. However, there is a lack of robust techniques and affordable equipment for avian embryo detection and sexual segregation at the early stages. In this work, we aimed to evaluate the potential use of thermal micro cameras for detecting embryos in quail eggs via thermal images during the first 168 h (7 days) of incubation. We propose a methodology to collect data during the incubation period. Additionally, to support the visual analysis, YOLO deep learning object detection algorithms were applied to detect unfertilized eggs; the results showed its potential to distinguish fertilized eggs from unfertilized eggs during the incubation period, after filtering radiometric images. We compared YOLOv4, YOLOv5 and SSD-MobileNet V2 trained models. The mAP@0.50 of the YOLOv4, YOLOv5 and SSD-MobileNet V2 was 98.62%, 99.5% and 91.8%, respectively. We also compared three testing datasets for different intervals of rotation of eggs, as our hypothesis was that fewer turning periods could improve the visualization of fertilized egg features, and applied three treatments: 1.5 h, 6 h, and 12 h. The results showed that turning eggs in different periods did not exhibit a linear relation, as the F1 Score for YOLOv4 of detection for the 12 h period was 0.569, that for the 6 h period was 0.404 and that for the 1.5 h period was 0.384. YOLOv5 F1 Scores for 12 h, 6 h and 1.5 h were 1, 0.545 and 0.386, respectively. SSD-MobileNet V2 performed F1 scores of 0.60 for 12 h, 0.22 for 6 h and 0 for 1.5 h turning periods.

Developing countries in Asia widely use manual seed broadcasting methods due to a lack of appropriate seeding machinery. The agricultural sector is currently facing labor shortages and high labor costs, especially seasonal labor shortages for broadcasting and transplanting operations. However, the primary constraint in adopting existing broadcasting seeders for small-scale farmers in developing countries is the high initial purchase costs. Therefore, developing locally commercial accessible technology for small-scale farmers is an urgent requirement. In this regard, attempt was taken to develop a new low-cost 3D printed seeder that can be used for multi-crop seed broadcasting operations when integrated with an autonomous terrain vehicle. A new seed metering mechanism was proposed for seed broadcasting that can be controlled electronically from the autonomous terrain vehicle. Positional sensors based on the real time kinematics—global navigation satellite system (RTK-GNSS) were used to record positional information. The best observation was noted at a vehicle operational speed of 0.351 ms−1 and had a coefficient of variation (CV) referring to the distribution uniformity of seeds of 19% for green peas, 22% for cowpeas, and 25% for chickpeas. The developed seeder could spread multi-crop seeds and adjust the seed rates electronically at the different ranges of rotational speeds. Therefore, the use of 3D printed fabricated prototype seed broadcasting units with small-scale autonomous vehicles can be implemented to help in labor supplements and perform the broadcasting of different seeds.

No abstract

No abstract