Search

Horticultural products such as fruits, vegetables, and tea offer a range of important nutrients such as protein, carbohydrates, vitamins and lipids. However, the present yield and quality do not meet the requirements of the rapid population growth associated with global climate change, the decline in horticultural practitioners, poor automation, and epidemic diseases such as COVID-19. In this context, smart horticulture is expected to greatly improve the land output rates, resource-use efficiency, and productivity, all of which should facilitate the sustainable development of the horticulture industry. Emerging technologies, such as artificial intelligence, big data, the Internet of Things, and cloud computing, play an important role. This paper reviews past developments and current challenges, offering future perspectives for horticultural chain management. We expect that the horticulture industry would benefit from integration with smart technologies. This requires the use of novel solutions to build a new advanced system encompassing smart breeding, smart cultivation, smart transportation, and smart sales. Finally, a new development approach combining precise perception, smart operation, and smart control should be instituted in the horticulture industry. Within 30 years, we expect that the industry will embrace mechanical, automatic, and informational production to transform into a smart industry.

Objective: Based on the inhibitory effect of bacillomycin D (BD) and chitosan (CTS) on Penicillium digitatum, the effect of BD and CTS combination on post-harvest disease control and freshness preservation of sugar oranges was investigated. Method: Sugar oranges as the research object, mold strains were screened in the moldy sugar oranges for identification, and the minimum inhibitory concentration and FIC index of BD and CTS against P. digitatum were determined by the checkerboard microdilution method. Finally, treated sugar oranges with BD and CTS were stored and preserved under the conditions of 28 ℃ and 90% relative humidity, and the changes of morbidity, weight loss, hardness, soluble solids, enzyme activity and other indicators were determined. Result: The isolated and purified fungus from the sugar orange was identified as P. digitatum and named P. digitatum strain L6. The minimum inhibitory concentrations of BD and CTS for P. digitatum were 42.19 and 62.5 mg/L, respectively, and the FIC index was 0.75, indicating that the inhibition of BD and CTS on P. digitatum was cumulative. The storage and preservation results showed that after 15 days of storage, the weight loss rates of BD, CTS, and BD+CTS treated sugar oranges were 0.99%, 1.62%, and 1.09%, respectively, which were 66.44%, 45.08%, and 63.05% lower compared to the control groups. In addition, the results of changes in titratable acidity, soluble solids, ascorbic acid, and plant-related enzyme activities also indicated that BD, CTS, and BD+CTS combination treatments had good preservation effects on sugar oranges, with the BD+CTS combination treatment group having the best results. Conclusion: Both BD and CTS had significant inhibition effect on P. digitatum, which could be used for storage and preservation of sugar orange to extend the shelf life. The combination of the two could not only enhance the preservation effect, but also the low price of CTS could reduce the use of BD and save the cost.

Abstract This study aimed to perform the first external validation of the modified Child-Turcotte-Pugh score based on plasma ammonia (aCTP) and compare it with other risk scoring systems to predict survival in patients with cirrhosis after transjugular intrahepatic portosystemic shunt (TIPS) placement. We retrospectively reviewed 473 patients from three cohorts between January 2016 and June 2022 and compared the aCTP score with the Child-Turcotte-Pugh (CTP) score, albumin-bilirubin (ALBI), model for end-stage liver disease (MELD) and sodium MELD (MELD-Na) in predicting transplant-free survival by the concordance index (C-index), area under the receiver operating characteristic curve, calibration plot, and decision curve analysis (DCA) curve. The median follow-up time was 29 months, during which a total of 62 (20.74%) patients died or underwent liver transplantation. The survival curves for the three aCTP grades differed significantly. Patients with aCTP grade C had a shorter expected lifespan than patients with aCTP grades A and B (P

In this study, we propose a high-throughput and low-cost automatic detection method based on deep learning to replace the inefficient manual counting of rapeseed siliques. First, a video is captured with a smartphone around the rapeseed plants in the silique stage. Feature point detection and matching based on SIFT operators are applied to the extracted video frames, and sparse point clouds are recovered using epipolar geometry and triangulation principles. The depth map is obtained by calculating the disparity of the matched images, and the dense point cloud is fused. The plant model of the whole rapeseed plant in the silique stage is reconstructed based on the structure-from-motion (SfM) algorithm, and the background is removed by using the passthrough filter. The downsampled 3D point cloud data is processed by the DGCNN network, and the point cloud is divided into two categories: sparse rapeseed canopy siliques and rapeseed stems. The sparse canopy siliques are then segmented from the original whole rapeseed siliques point cloud using the sparse-dense point cloud mapping method, which can effectively save running time and improve efficiency. Finally, Euclidean clustering segmentation is performed on the rapeseed canopy siliques, and the RANSAC algorithm is used to perform line segmentation on the connected siliques after clustering, obtaining the three-dimensional spatial position of each silique and counting the number of siliques. The proposed method was applied to identify 1457 siliques from 12 rapeseed plants, and the experimental results showed a recognition accuracy greater than 97.80%. The proposed method achieved good results in rapeseed silique recognition and provided a useful example for the application of deep learning networks in dense 3D point cloud segmentation.

Timely harvest can effectively guarantee the yield and quality of rapeseed. In order to change the artificial experience model in the monitoring of rapeseed harvest period, an intelligent prediction method of harvest period based on deep learning network was proposed. Three varieties of field rapeseed in the harvest period were divided into 15 plots, and mobile phones were used to capture images of silique and stalk and manually measure the yield. The daily yield was divided into three grades of more than 90%, 70–90%, and less than 70%, according to the proportion of the maximum yield of varieties. The high-dimensional features of rapeseed canopy images were extracted using CNN networks in the HSV space that were significantly related to the maturity of the rapeseed, and the seven color features of rapeseed stalks were screened using random forests in the three color-spaces of RGB/HSV/YCbCr to form a canopy-stalk joint feature as input to the subsequent classifier. Considering that the rapeseed ripening process is a continuous time series, the LSTM network was used to establish the rapeseed yield classification prediction model. The experimental results showed that Inception v3 of the five CNN networks has the highest prediction accuracy. The recognition rate was 91% when only canopy image features were used, and the recognition rate using canopy-stalk combined features reached 96%. This method can accurately predict the yield level of rapeseed in the mature stage by only using a mobile phone to take a color image, and it is expected to become an intelligent tool for rapeseed production.

The resolution limitation is a main problem in the time-reversed field shaping system. In this paper, a subwavelength field shaping approach is proposed by combining time reversal (TR) technique and defective resonant metasurfaces. The subwavelength 2D and 3D field patterns can be effectively realized in both single-target and multi-target field-shaping systems. The resolution in both single-target and multi-target field-shaping systems aided with defective metasurfaces is approximately 0.1 λ, which is corresponding to the dimension of the applied metasurface cell. The proposed technique can be used for wireless power transfer, industrial heating, and localized wireless communications.

Rootstock grafting is an important method to improve the yield and quality of seedlings. Pumpkin is the rootstock of watermelon, melon, and cucumber, and the root phenotype of rootstock is an important reference for breeding. At present, the root phenotype is mainly measured by scanners, with which it is difficult to achieve non-destructive and in situ measurements. In this work, we propose a method for non-destructive measurement of the root phenotype on the surface layer of the root ball of pumpkin rootstock plug seedlings and an accurate estimation of the surface area, length, and volume of total root using an AZURE KINECT sensor. Firstly, the KINECT is used to capture four-view color and depth images of the root surface, and then multi-view images are spliced to obtain a complete image of the root surface. After preprocessing of the images, we extract the roots from the root ball. For root phenotype measurements, the surface areas of the surface roots and root ball are calculated, followed by calculating root encapsulation. Next, the non-overlapping roots in the surface root image are extracted, and the ratio of the surface area to the skeleton length is used as the average diameter of total root. Based on the high correlation between the surface area of surface root and the surface area of total root, a linear fitting model is established to estimate the surface area, length, and volume of total root. The experiment ultimately showed that the measurement error for the average diameter of total root is less than 30 μm, and consistency with the scanner is higher than 93.3%. The accuracy of the surface area of total root estimation was found to be more than 88.1%, and the accuracy of the root length of total root estimation was observed to be greater than 87.2%. The method proposed in this paper offers similar accuracy to a scanner, which meets the needs of non-destructive root phenotype research. This method is expected to replace root scanners for high-throughput phenotypic measurements and provides a new avenue for root phenotype measurements of pumpkin rootstocks. This technology will provide key basic data for evaluating the root growth of pumpkin rootstocks.