Search

Soil salinization is one of the important factors limiting the sustainable development of agriculture and causing land degradation. In order to explore the effect of growth stage division on the estimation accuracy of soil salt content (SSC) in fields cropped with sunflower, multiple time‐series unmanned aerial vehicle (UAV) were used to monitor SSC in the Hetao Irrigation District, Inner Mongolia, China. Aerial and field campaigns for four growth stages of sunflowers in six study areas were conducted from July to September in 2021. The ground samplings of electrical conductivity (EC), leaf area index (LAI), plant height (H), and leaf chlorophyll content (CHL) were taken simultaneously with (UAV) multispectral images. The correlation between six vegetation indices (VIs), four salinity indices (SIs), three crop parameters (LAI, CHL, H) and SSC was investigated. The optimal parameters were determined and used as input variables to establish SSC estimation model using artificial neural network (ANN), random Forest (RF), multiple linear regression (MLR) algorithm, respectively. The results show that the division of growth stages could improve the correlation between spectral index, growth parameters and SSC, and the estimation model by each growth stage was more accurate than that of the whole growth period. Among the spectral indices, the VIs showed a higher correlation with SSC than the SIs. Among the crop parameters, the LAI was the most sensitive to the degree of soil salinization. The nonlinear regression algorithm (ANN, RF) performed better than the linear regression model (MLR) in the application of SSC estimation, and the best estimation models for the four growth stages of sunflower were the ANN_SSC models. This study proposed a fast and low‐cost method to monitor the soil salinization of sunflower‐cropped fields in time‐series and provided a reference for the quick perception and prevention of soil salinization. [ABSTRACT FROM AUTHOR]

Timely and accurate estimation of crop evapotranspiration (ET c) is essential for efficient irrigation management at the farmland scale. However, effective decision-making for irrigation scheduling requires high spatiotemporal-resolution data to provide within-field heterogeneity information. The objective of this study was to evaluate the use of optical and thermal information obtained from an unmanned aerial vehicle (UAV) to quantify maize (Zea mays L.) ET c within the framework of the FAO-56 dual crop coefficient approach. Canopy temperature-based crop water stress index (CWSI) and number of degrees above canopy threshold (DACT) were used to determine the stress coefficient (K s). Three types of normalized difference vegetation index (NDVI), soil adjusted vegetation index (SAVI), and enhanced vegetation index (EVI) were adopted to determine basal crop coefficient (K cb). Then the two forms of stress coefficient in combination with three vegetation indices (VIs) were evaluated to estimate daily crop ET c under different irrigation treatments over two years in the southwest region of Inner Mongolia, China. The results demonstrated that the combination of NDVI and CWSI produced the best estimates of maize ET c , with R2 of 0.84 and root mean square error (RMSE) of 0.50 mm/day. The DACT-based model also performed well, with R2 of 0.77–0.80 and RMSE of 0.53 mm/day. Although the results varied with irrigation levels, the daily mean bias error of ET c predictions over different years indicated acceptable accuracy, with a mean bias error (MBE) of 0.48 mm/day. Sensitivity analyses indicated that ET c models were most sensitive to the slope of the linear regression between K cb and VIs, followed by soil evaporation constant and influential factor in K s_DACT. From this study, the combinations of vegetation index and CWSI, as well as DACT, were recommended as alternative approaches for estimating ET c due to intrinsic simplicity and easy interpretation. These ET c models relying on UAV-based multi-sensor data thus show promising potential in farmland-scale applications. • The potential of low-cost UAV-based image data for maize ET c prediction was evaluated. • Three vegetation indices NDVI, SAVI, and EVI were used to determine K cb. • Two different forms of K s methods (CWSI and DACT) were compared. • The robustness and adaptability of UAV-based ET c estimation model across different irrigation treatments was discussed. [ABSTRACT FROM AUTHOR]

• The effect of imagery collected platform on the fixed-threshold method was assessed. • Machine learning regression algorithms were adopted to relate UAV VIs to FVC. • The established FVC estimation model was tested for different growing season. • RF model performed well for different growing seasons, stages, and water status. Crop water stress is an inevitable and increasing challenge for agriculture. To improve crop water use efficiency, management of water stress and accurate estimation of crop traits were required. Among crop traits used to detect crop growth status and predict yield, fractional vegetation cover (FVC) is of great significance. We conducted studies in a maize field located in Inner Mongolia, China with different irrigation levels during 2018 and 2019 growing seasons. UAV RGB imagery was captured to investigate the effect of image sensors on thresholds obtained by the fixed-threshold method (proposed in our recent study), and to provide reference FVC (FVC UAV_R) for FVC models based on UAV multispectral imagery. Five vegetation indices (VIs), calculated from UAV multispectral imagery, and three regression algorithms (RF: random forest, ANN: artificial neural network, and MLR: multivariate linear regression) were used to build the FVC model suitable for different growing seasons, growth stages, and crop water stress. The results showed that there was a change in thresholds obtained using the fixed-threshold method based on different image sensors, but this change did not make a big difference on the accuracy of FVC UAV_R , with the R2 difference of 0.01 and the RMSE difference of 0.01. As for the three FVC regression models, RF model was the most suitable model when these models established in 2018 were used to estimate maize FVC in 2019 for different growth stages and water stress. The low estimation accuracy for high FVC levels was the reason why MLR model could not be used in the other maize growing season. This study provides a low cost and easy way to estimate maize FVC and its inter-field variability under various water status in different maize growing seasons or growth stages. [ABSTRACT FROM AUTHOR]

Different irrigation treatments in plants significantly influence CO 2 emissions. This study investigated the net ecosystem CO 2 exchange (NEE) and yield of maize cropland under different irrigation treatments to determine the optimal irrigation treatment at different maize growth stages. Different degrees of water deficit irrigation were carried out in four fields: full irrigation, 100% evapotranspiration (ET); moderate deficit irrigation, 65% ET; and severe deficit irrigation, 40% ET. Automated static and dynamic chambers were used to measure CO 2 fluxes. We found that NEE showed a "V" -shaped trend during the day and slowly declined at night. The light response parameters were calculated using the rectangular hyperbola model, and the leaf area index accounted for 85% of the light use efficiency. In the late vegetative stage, severe deficit irrigation led to a reduction in CO 2 emissions but significantly reduced the crop growth rate (p < 0.05). In the reproductive stage, full irrigation not only reduced CO 2 emissions but also significantly increased the crop growth rate. In the maturation stage, moderate deficit irrigation reduced CO 2 emissions, and the crop growth rate decreased compared with full irrigation, but the decrease was not significant (p > 0.05). Compared with full irrigation throughout the growth period, full irrigation in the reproductive stage and moderate deficit irrigation in the late vegetative and reproductive stages both increased the maize carbon sink capacity and light use efficiency at higher leaf area index, and the harvest index did not decrease significantly. However, inadequate irrigation in the reproductive stage or severe deficit irrigation in the late vegetative and maturation stages significantly reduced the maize harvest index. Thus, full irrigation during the reproductive stage, and moderate deficit irrigation during the late vegetative and reproductive stages were the optimal irrigation strategies for saving water, reducing CO 2 emissions and ensuring a satisfactory maize yield. • NEE in maize cropland under different irrigation treatments in semiarid Inner Mongolia. • Automated static and dynamic chambers were used to measure night and day maize NEE. • Full irrigation in reproductive stage can decrease CO 2 emission. • Moderate deficit irrigation in late vegetative and maturation stages reduced CO 2 emission. • Reasonable irrigation can reduce CO 2 emission and maintain high harvest index. [ABSTRACT FROM AUTHOR]

An improved semantic segmentation method based on object contextual representations network (OCRNet) is proposed to accurately identify zucchinis intercropped with sunflowers from unmanned aerial vehicle (UAV) visible images taken over Hetao Irrigation District, Inner Mongolia, China. The proposed method improves on the performance of OCRNet in two respects. First, based on the object region context extraction structure of the OCRNet, a branch that uses the channel attention module was added in parallel to rationally use channel feature maps with different weights and reduce the noise of invalid channel features. Secondly, Lovász-Softmax loss was introduced to improve the accuracy of the object region representation in the OCRNet and optimize the final segmentation result at the object level. We compared the proposed method with extant advanced semantic segmentation methods (PSPNet, DeepLabV3+, DNLNet, and OCRNet) in two test areas to test its effectiveness. The results showed that the proposed method achieved the best semantic segmentation effect in the two test areas. More specifically, our method performed better in processing image details, segmenting field edges, and identifying intercropping fields. The proposed method has significant advantages for crop classification and intercropping recognition based on UAV visible images, and these advantages are more substantive in object-level evaluation metrics (mIoU and intercropping IoU). [ABSTRACT FROM AUTHOR]

• Ground measurements of canopy structure and chlorophyll content were adopted. • The annual change of relationships between UAV VIs and Gs was explored. • Non-linear regression algorithms were adopted to relate UAV VIs to Gs. • Multiple linear regression approach performed well and was considered very robust. To further assess the sensitivity of crop chlorophyll and structure based on UAV vegetation indices (VIs) to maize water stress, a study was carried out in a maize field located in Inner Mongolia, China, with various levels of deficit irrigation over the entire 2018 and 2019 growing seasons. Ground measurements of stomatal conductance (Gs), leaf area index and leaf chlorophyll were used as references for maize water status, canopy structure and chlorophyll content, respectively. Four structure VIs and two chlorophyll VIs, and three regression algorithms (multiple linear, random forest and artificial neural networks regression) were adopted. The results showed that canopy structure derived from VIs had a significant correlation (p < 0.001) with Gs with the highest r value of 0.64 (n = 270) in 2018 and 2019. The transformed chlorophyll absorption in reflectance index, chlorophyll VI, could only estimate severe maize water stress with an r value of −0.47 (p < 0.001, n = 270) for the drier 2019. The water stress sensitivity of chlorophyll and structure VIs maybe significantly influenced by different responses of canopy structure and chlorophyll concentration to water stress, and the different spectral resolution of UAV multispectral cameras. Compared to non-linear machine learning regression algorithms, the multiple linear regression was robust enough to relate UAV-based multispectral VIs to Gs with coefficients of determination of 0.48 and 0.45 (n = 270) for 2018 and 2019, respectively. Although stable significant correlations were found between UAV multispectral VIs and Gs, annual changes in these specific expressions were also observed. Overall, our results demonstrated the potential of using structure VIs derived from UAV multispectral images and multiple linear regression approach to estimate maize water status in field scale. [ABSTRACT FROM AUTHOR]

The rapid, accurate, and real-time estimation of crop coefficients at the farm scale is one of the key prerequisites in precision agricultural water management. This study aimed to map the maize crop coefficient (Kc) with improved accuracy under different levels of deficit irrigation. The proposed method for estimating the Kc is based on multispectral images of high spatial resolution taken using an unmanned aerial vehicle (UAV). The analysis was performed on five experimental plots using Kc values measured from the daily soil water balance in Ordos, Inner Mongolia, China. To accurately estimate the Kc, the fraction of vegetation cover (fc) derived from the normalized difference vegetation index (NDVI) was used to compare with field measurements, and the stress coefficients (Ks) calculated from two vegetation index (VI) regression models were compared. The results showed that the NDVI values under different levels of deficit irrigation had no significant difference in the reproductive stage but changed significantly in the maturation stage, with a decrease of 0.09 with 72% water applied difference. The fc calculated from the NDVI had a high correlation with field measurement data, with a coefficient of determination (R2) of 0.93. The ratios of transformed chlorophyll absorption in reflectance index (TCARI) to renormalized difference vegetation index (RDVI) and TCARI to soil-adjusted vegetation index (SAVI) were used, respectively, to establish two types of Ks regression models to retrieve Kc. Compared to the TCARI/SAVI model, the TCARI/RDVI model under different levels of deficit irrigation had better correlation with Kc, with R2 and root-mean-square error (RMSE) values ranging from 0.68 to 0.80 and from 0.140 to 0.232, respectively. Compared to Kc calculated from on-site measurements, the Kc values retrieved from the VI regression models established in this study had greater ability to assess the field variability of soil and crops. Overall, use of the UAV-measured multispectral vegetation index approach could improve water management at the farm scale. [ABSTRACT FROM AUTHOR]

With increasing consumption, plastic mulch benefits agriculture by promoting crop quality and yield, but the environmental and soil pollution is becoming increasingly serious. Therefore, research on the monitoring of plastic mulched farmland (PMF) has received increasing attention. Plastic mulched farmland in unmanned aerial vehicle (UAV) remote images due to the high resolution, shows a prominent spatial pattern, which brings difficulties to the task of monitoring PMF. In this paper, through a comparison between two deep semantic segmentation methods, SegNet and fully convolutional networks (FCN), and a traditional classification method, Support Vector Machine (SVM), we propose an end-to-end deep-learning method aimed at accurately recognizing PMF for UAV remote sensing images from Hetao Irrigation District, Inner Mongolia, China. After experiments with single-band, three-band and six-band image data, we found that deep semantic segmentation models built via single-band data which only use the texture pattern of PMF can identify it well; for example, SegNet reaching the highest accuracy of 88.68% in a 900 nm band. Furthermore, with three visual bands and six-band data (3 visible bands and 3 near-infrared bands), deep semantic segmentation models combining the texture and spectral features further improve the accuracy of PMF identification, whereas six-band data obtains an optimal performance for FCN and SegNet. In addition, deep semantic segmentation methods, FCN and SegNet, due to their strong feature extraction capability and direct pixel classification, clearly outperform the traditional SVM method in precision and speed. Among three classification methods, SegNet model built on three-band and six-band data obtains the optimal average accuracy of 89.62% and 90.6%, respectively. Therefore, the proposed deep semantic segmentation model, when tested against the traditional classification method, provides a promising path for mapping PMF in UAV remote sensing images. [ABSTRACT FROM AUTHOR]

• Redundant information of the imagery was removed by the feature selection method. • U-Net algorithm is more suitable for crop mapping in the irrigation district. • The key early window period of major crop mapping was determined. • The crop mapping models with strong versatility in the early period were obtained. The information on spatial distribution pattern and area of different crops is particularly useful for monitoring and managing the sustainability of agricultural resources. However, there are still challenges in timely mapping of crop types and planting areas to support production during the present growing season. Here, we proposed to use time-series Sentinel-2 imagery and a deep learning method for major crop mapping (wheat, maize, sunflower, and squash) in the agricultural production areas in the Hetao Irrigation District (HID), Inner Mongolia Autonomous Region, China. A feature selection method that combines the global separability index and feature recursive elimination was proposed to optimize the band features of Sentinel-2 imagery. The Random Forest, Extreme Gradient Boosting, U-Net and deeplabv3+ algorithms were then used for image classification with all spectral bands and the selected band features, respectively. The selected bands and best method were used to determine the key time window for the early identification of four major crops. Integrate all image datasets in the early window period, and build mapping models suitable for images at different times in the window period by combining the selected bands and best method. The results showed that the proposed feature selection method effectively removed redundant features and improved the classification efficiency. The U-Net algorithm was more suitable for the classification of major crops in HID with a higher accuracy whether using all band features or the selected features. The key periods for early recognition of wheat, maize, sunflower, and squash in HID were mid-May, late June, early July, and late June, respectively. The constructed models for crop type mapping in the early period were tested in 2020 and 2021 in the HID, and the F1-score, the mean intersection-over-union and overall accuracy were 86.71 %, 84.66 %, and 98.89 %, respectively. The proposed feature selection method, U-Net classification algorithm, and screening of the early mapping window period in this study provide an efficient, accurate and timely crop mapping tool in HID. Future research can further verify the applicability of the early identification time window proposed in this study to support other agricultural management activities in the areas. [ABSTRACT FROM AUTHOR]