Deep-learning-based object detection algorithms have significantly improved the performance of wheat spike detection. However, UAV images crowned with small-sized, highly dense, and overlapping spikes cause the accuracy to decrease for detection. This paper proposes an improved YOLOv5 (You Look Only Once)-based method to detect wheat spikes accurately in UAV images and solve spike error detection and miss detection caused by occlusion conditions. The proposed method introduces data cleaning and data augmentation to improve the generalization ability of the detection network. The network is rebuilt by adding a microscale detection layer, setting prior anchor boxes, and adapting the confidence loss function of the detection layer based on the IoU (Intersection over Union). These refinements improve the feature extraction for small-sized wheat spikes and lead to better detection accuracy. With the confidence weights, the detection boxes in multiresolution images are fused to increase the accuracy under occlusion conditions. The result shows that the proposed method is better than the existing object detection algorithms, such as Faster RCNN, Single Shot MultiBox Detector (SSD), RetinaNet, and standard YOLOv5. The average accuracy (AP) of wheat spike detection in UAV images is 94.1%, which is 10.8% higher than the standard YOLOv5. Thus, the proposed method is a practical way to handle the spike detection in complex field scenarios and provide technical references for field-level wheat phenotype monitoring.

To improve the efficiency and effectiveness of mosaicking unmanned aerial vehicle (UAV) images, we propose in this paper a rapid mosaicking method based on scale-invariant feature transform (SIFT) for mosaicking UAV images used for crop growth monitoring. The proposed method dynamically sets the appropriate contrast threshold in the difference of Gaussian (DOG) scale-space according to the contrast characteristics of UAV images used for crop growth monitoring. Therefore, this method adjusts and optimizes the number of matched feature point pairs in UAV images and increases the mosaicking efficiency. Meanwhile, based on the relative location relationship of UAV images used for crop growth monitoring, the random sample consensus (RANSAC) algorithm is integrated to eliminate the influence of mismatched point pairs in UAV images on mosaicking and to keep the accuracy and quality of mosaicking. Mosaicking experiments were conducted by setting three types of UAV images in crop growth monitoring: visible, near-infrared, and thermal infrared. The results indicate that compared to the standard SIFT algorithm and frequently used commercial mosaicking software, the method proposed here significantly improves the applicability, efficiency, and accuracy of mosaicking UAV images in crop growth monitoring. In comparison with image mosaicking based on the standard SIFT algorithm, the time efficiency of the proposed method is higher by 30%, and its structural similarity index of mosaicking accuracy is about 0.9. Meanwhile, the approach successfully mosaics low-resolution UAV images used for crop growth monitoring and improves the applicability of the SIFT algorithm, providing a technical reference for UAV application used for crop growth and phenotypic monitoring.

The impact of global warming on crop growth periods and yields has been evaluated by using crop models, which need to provide various kinds of input datasets and estimate numerous parameters before simulation. Direct studies on the changes of climatic factors on the observed crop growth and yield could provide a more simple and intuitive way for assessing the impact of climate change on crop production. In this study, four cultivars which were planted over more than 15 years in eight test stations in the Northern Winter Wheat Region of China were selected to investigate the relationships between growth periods, grain yields, yield components and temperatures. It was found that average temperatures and heat degree-days (HDD) during the winter wheat growing seasons tended to increase over time series at most study sites. The length of growth period and growing degree days (GDD) were not fixed for a given cultivar among different years and locations, and the variation on the periods from sowing to jointing was relatively greater than in the other periods. The increasing temperature mainly shortened the periods from sowing to jointing and jointing to anthesis, which led to the decrease in entire growth periods. Positive relationships between spike number, grain number per spike, grain yields and average temperatures were identified in the Northern Winter Wheat Region of China. The grain yield in the study area increased by 406.3 kg ha⁻¹ for each 1 °C increase in average temperature. Further, although the positive relationship between grain yield and HDD was found in our study, the heat stress did not lead to the wheat yield decline in the study region. Temperature is a major determinant of wheat growth and development, the average temperature and the frequency of heat stresses are projected to increase in the future, so understanding the effect of temperature on wheat production and adopting appropriate adaptation are required for the implementation of food security policies.

Sowing and transplanting dates are important cultivation factors for rice production. Therefore, the present study focused on developing a rice model system that would be able to determine sowing and transplanting dates for diverse cultivars and planting methods in different agro-ecological zones. Different model parameters were integrated into a rice model system on the basis of their interaction effects in this study. The results showed that sowing and transplanting dates designed by the rice model system were approached to the planting dates recommended by local agricultural experts for high yield practices, with root mean squared error (RMSE) of 5.3 to 14.74 days. The model system accurately simulated suitable sowing and transplanting dates under most scenarios with relatively low RMSE, high linear correlation coefficient (R²), and model efficiency (EF). Using the model system recommendations, rice yield under manual transplanting in low fertility soil was increased the most (5.5%), while for direct sowing in high fertility soil, yield increase was modest (0.8%). The newly-developed rice model system can act as a technical approach to design suitable sowing and transplanting dates for achieving high yield and effective crop production.

The crop model based on physiology and ecology has been widely applied to the simulation of regional potential productivity. By determining the appropriate spatial resolution of meteorological data required for model simulation for different regions, we can reduce the difficulty of acquiring model input data, thereby improving the regional computing efficiency of the model and increasing the model applications. In this study, we investigated the appropriate spatial resolution of meteorological data needed for the regional potential productivity simulation of the WheatGrow model by scale effect index and verify the feasibility of using the landform to obtain the appropriate spatial resolution of meteorological data required by the potential productivity simulation for the winter wheat region of China. The research results indicated that the spatial variation of landforms in the winter wheat region of China is significantly correlated to the spatial variation of multi-year meteorological data. Based on the scale effect index, we can obtain a spatial distribution of appropriate spatial resolution for the meteorological data required for the regional potential productivity simulation of the WheatGrow model for the winter wheat region of China. Moreover, although we can use the spatial heterogeneity of landforms to guide the selection of appropriate spatial resolution for the meteorological data, in the regions where the spatial heterogeneity of the landform is relatively weak or relatively strong over a small range, the method of using a single heterogeneity index derived from semi-variogram cannot well reflect the scale effect of simulation results and needs further improvement.