Recognition and localization of strawberries from 3D binocular cameras for a strawberry picking robot using coupled YOLO/Mask R-CNN

In order to detect kiwifruit quickly and accurately in orchard environments for the picking robot, this paper proposed a detection method based on a lightweight YOLOv4-GhostNet network. The implementations of the method are as follows: The original CSP-Darknet53 backbone network model was replaced by GhostNet, a feature layer facilitating small object detection was introduced in the feature fusion layer, and part of the ordinary convolution was replaced by a combination of 1 × 1 convolution and depth-separable convolution to reduce the computational pressure caused by the fused feature layer. The parameters of the new network are reduced, and the generalization ability of the model is improved by loading pre-training weights and freezing some layers. The trained model was tested, and the results showed that the detection performances were better than that of the original YOLOv4 network. The F1 value, map, and precision were improved on the test set, which were 92%, 93.07%, and 90.62%, respectively. The size of weight parameters was reduced to 1/6 of the original YOLOv4 network, and the detection speed reached 53 FPS. Therefore, the method proposed in this study shows the features of fast recognition, lightweight parameters, and high recognition accuracy, which can provide technical support for vision systems of kiwifruit picking robots.

Here, the improved multi‐scale YOLO algorithm (Improved‐YOLOv3) is presented, which was proposed to realize fast and accurate recognition of citrus fruit in a field environment. With the modification of the YOLO‐styled network model, a darknet‐53 backbone network with residual modules was designed. A multi‐scale detection module was to construct a network model for rapid recognition of citrus fruit in complex environments. Using the improved model to detect and identify citrus fruit targets, the network model can extract more feature information. The improved YOLOv3 model was tested with citrus data, and the detection performance of the improved network and the influence of the number of backbone network layers on the feature extraction capability were compared. The results showed a good detection ability (detection rate, accuracy, map, detection speed) for the target fruit, and the improved YOLOv3 network showed higher accuracy. Moreover, the performance of different training models were compared: the Improved‐YOLOv3 has stronger robustness, higher detection accuracy, and shorter training time, and can recognize citrus in complex field environments. The experiment showed that the precision of Improved‐YOLOv3 was 90.5% and the accuracy reached 94.3%, the recall rate was 90.3%, the detection time was 9.89 ms per frame, which could provide technical support for this visual recognition system of citrus picking robot.

Sweetness or sugar content, represented by soluble solids contents (SSC), is a vital quality trait in watermelon breeding which can be assessed by the refractive index method. However, sampling watermelon juice out of the pulp is a process that is both labor-intensive and error-prone. In this study, we developed an automatic SSC measurement system for watermelon breeding to improve efficiency and decrease costs. First, we built an automatic cutting system to cut watermelons into precise halves, in which a laser rangefinder is used to measure the distance from the upper surface of the watermelon to itself, and thus, the diameter is estimated. The experiments showed a high correlation between the estimated diameters and the ground truths, with and . Then, we built an automatic Brix measurement system to obtain the Brix data from a central point on the watermelon’s section, where an image analysis procedure is applied to locate the testing point. This is then transformed to the camera coordination system, and a refractometer is driven by a 3-axis robotic arm to reach the testing point. Brix measurement experiments were conducted using three vertical gaps and four lateral gaps between the probe of the refractometer and the pulp. The result showed that the best parameters were a vertical gap of 4 mm and a lateral gap of 2 mm. The average accuracy reached 98.74%, which indicates that this study has the potential to support watermelon breeding research.

Haploid breeding technology has advantages in terms of saving time and reducing labor intensity and costs. However, the low induction rate limits the application of this technology. Previous researchers found that heat shock can increase the rate of Embryo-like structures (ELSs) induction. However, molecular mechanisms underlying heat-shocked haploid induction remain poorly understood. In the current study, unfertilized ovules of watermelon were subjected to heat shock for 0–5 days and conducted transcriptomics sequencing and DIA-based proteomics sequencing. Results indicated that, in contrast to the non-heat-shock condition, the expression level of protein phosphatase 2C (PP2C), a negative regulator in abscisic acid (ABA) signal transduction pathway, was repressed, and the expression level of Sucrose-non-fermenting 1-related protein kinases (SnRK2) was activated. The activated SnRK2s are enabled to promote the accumulation of storage substances in ovules. Through analysis, the expression of many genes involved in the biosynthesis of unsaturated fatty acids and amino acids has indeed been upregulated. In conclusion, our findings demonstrate that heat shock promotes the accumulation of storage substances in unfertilized ovules by activating the signal transduction process of ABA, which correspondingly increases ELSs induction rate.

Improving plant regeneration ability and shortening regeneration time can promote the development of genetic transformation breeding technology for horticultural crops. We optimized several culture conditions, including explant type, mother plant genotype, and medium, to improve shoot formation in winter squash (Cucurbita maxima Duch.). Histological analysis of the occurrence of shoots was also carried out. The results indicate that cotyledon was the most suitable explant for inducing the shoot regeneration of winter squash. We found that ‘Jin-li’ had a shorter shoot induction time and a higher average number of shoots. The highest induction rate of 95.23% among the five lines. The average shoot induction rate of five lines was the highest (84.85%) on Murashige and Skoog (MS) medium supplemented with 2.0 mg/L 6-benzylaminopurine (6-BA) and 0.2 mg/L indole-3-acetic acid (IAA). We also found that there was an interaction between genotypes and induction media, and their interaction had a greater impact on the shoot induction rate than individual effects. Histological observation revealed that the induced shoots of winter squash cotyledons originated from subepidermal cells. We also found that the optimal medium for de novo root regeneration was 1/2 MS. We acclimatized and cultivated regenerated plants and harvested their fruits, which maintained the characteristics of mother plants. These findings lay an important foundation for further research on direct shoot regeneration and accelerate its application in winter squash genetic transformation.