C. Liu’s research while affiliated with Jiangsu University and other places

The aerial platforms have the potential of realizing the precision agriculture, especially the rational allocation of water resources and the intelligent irrigation management of crops. This paper investigates a Unmanned Aerial Vehicle (UAV)-assisted intelligent monitoring method of soil temperature and moisture. The UAV platform is equipped with a high-precision infrared sensor to sample the discrete-time images of the target ground object. By combining the adaptive median filtering, the mean filtering and the Canny edge detection algorithms, we design a composite image preprocessing scheme to obtain the precise location information, and then extract the crop canopy temperature. Specifically, the soil moisture prediction model based on Radial Basis Function Neural Network (RBFNN) and Principal Component Analysis (PCA) are established to achieve accurate prediction of farmland moisture. Experimental results show that (1) the proposed adaptive method is capable of improving the detection accuracy of the target canopy boundary and soil background; (2) by establishing the linear regression model between the real ground temperature and the UAV data, it is shown that the Canny algorithm can improve the extraction accuracy of the canopy temperature data from R^2 = 0.7673 to R^2 = 0.9355; (3) compared with the PCA method can greatly reduce the dimension of sample data, while the accuracy of the soil moisture content is almost unchanged. This article establishes a set of intelligent prediction methods for soil moisture content, which greatly improves the level of agricultural intelligence.

Water management is becoming a critical issue for sustainable agriculture, especially in the semi-arid region, where problems with water scarcity are rising. More accurate water status recovery in crops is required for precise irrigation through remote sensing technologies. These technologies have a lot of potential in intelligent irrigation because they allow for real-time environmental data collection. Nowadays, digital practices have been used, such as unmanned aerial vehicle (UAV), which plays an essential role in various applications related to crop management. Drones offer an exciting opportunity to track crop fields with high spatial and temporal resolution remote sensing to enhance water stress management in irrigation. Farmers have historically depended on soil moisture measurements and weather conditions to detect crop water status for irrigation scheduling. This review paper summarizes the use of UAV remote sensing data in crops for estimating the water status and gives a detailed summary of the potential capacity of UAV remote sensing for water stress application. The remote sensing techniques help modify agricultural practices to meet this significant challenge by providing repeated information on crop status at different scales and various performances during the season. UAVs successful implementation in water stress estimations depends on UAV features, such as flexibility of use in flight planning, low cost, reliability, autonomy, and capability of timely provision of high-resolution data. UAV with a thermal sensor is considered the most effective technique for detecting water stress using specific indices. Thermal imaging can identify water status variations and crop water stress index (CWSI). This CWSI acquired through UAV thermal sensors imagery can be acceptable for managing real-time irrigation to achieve optimum crop water efficiency.

This paper shows some practical experiences of using unmanned aerial vehicles-based platform for remote sensing in supporting precision agriculture mapping. There have been studies on unmanned aerial vehicles used to calculate plant water stress; however, the scientific reports of drone images that are used to predict best time and height are rare. The trial was conducted during 2020, in a five-year-old Anji tea plant experimental field, where drone captures images in a different time series of 27 flights during experimental days. This work aims to (1) investigate the appropriate thermography timing and altitude based on unmanned aerial vehicles remote sensing, (2) conduct a quantitative and qualitative study of various thermal orthomosaics and photographs, (3) establish workflow for high-resolution remote sensing application. All flights were operated at 3 m/s flying speed. Flights were performed during the testing day at about 09:00 h, 11:00 h, and 13:00 h. The drone images were taken at relative flying heights of 25 m, 40 m, and 60 m each day. The relationship between canopy temperature and plant-based variables was also established. The results reported that flights at 11:00 h and 60-m altitude orthomosaic could provide the best relation and accurate canopy temperature. On the other hand, the high relationship between stomatal conductance and canopy temperature was R2 0.98 at 11:00 h. The selection of optimal timing and altitude can provide rapid and reliable canopy temperature information. Overall, high resolution with low-altitude unmanned aerial vehicles images proved good relationship in order to assess the canopy temperature.