wGrapeUNIPD-DL: An open dataset for white grape bunch detection.

National and international Vitis variety catalogues can be used as image datasets for computer vision in viticulture. These databases archive ampelographic features and phenology of several grape varieties and plant structures images (e.g. leaf, bunch, shoots). Although these archives represent a potential database for computer vision in viticulture, plant structure images are acquired singularly and mostly not directly in the vineyard. Localization computer vision models would take advantage of multiple objects in the same image, allowing more efficient training. The present images and labels dataset was designed to overcome such limitations and provide suitable images for multiple cluster identification in white grape varieties. A group of 373 images were acquired from later view in vertical shoot position vineyards in six different Italian locations at different phenological stages. Images were then labelled in YOLO labelling format. The dataset was made available both in terms of images and labels. The real number of bunches counted in the field, and the number of bunches visible in the image (not covered by other vine structures) was recorded for a group of images in this dataset.

Radiative transfer model inversion using high-resolution hyperspectral airborne imagery - Retrieving maize LAI to access biomass and grain yield.

Mapping crop within-field yield variability provide an essential piece of information for precision agriculture applications. Leaf Area Index (LAI) is an important parameter that describes maize growth, vegetation structure, light absorption and subsequently maize biomass and grain yield (GY). The main goal for this study was to estimate maize biomass and GY through LAI retrieved from hyperspectral aerial images using a PROSAIL model inversion and compare its performance with biomass and GY estimations through simple vegetation index approaches. This study was conducted in two separate maize fields of 12 and 20 ha located in north-west Mexico. Both fields were cultivated with the same hybrid. One field was irrigated by a linear pivot and the other by a furrow irrigation system. Ground LAI data were collected at different crop growth stages followed by maize biomass and GY at the harvesting time. Through a weekly/biweekly airborne flight campaign, a total of 19 mosaics were acquired between both fields with a micro-hyperspectral Vis-NIR imaging sensor ranging from 400 to 850 nanometres (nm) at different crop growth stages. The PROSAIL model was calibrated and validated for retrieving maize LAI by simulating maize canopy spectral reflectance based on crop-specific parameters. The model was used to retrieve LAI from both fields and to subsequently estimate maize biomass and GY. Additionally, different vegetation indices were calculated from the aerial images to also estimate maize yield and compare the indices with PROSAIL based estimations. The PROSAIL validation to retrieve LAI from hyperspectral imagery showed a R2 value of 0.5 against ground LAI with RMSE of 0.8 m2/m2. Maize biomass and GY estimation based on NDRE showed the highest accuracies, followed by retrieved LAI, GNDVI and NDVI with R2 value of 0.81, 0.73, 0.73 and 0.65 for biomass, and 0.83, 0.69, 0.73 and 0.62 for GY estimation, respectively. Furthermore, the late vegetative growth stage at V16 was found to be the best stage for maize yield prediction for all studied indices.

Utilization of Landsat-8 data for the estimation of carrot and maize crop water footprint under the arid climate of Saudi Arabia.

The crop Water Footprint (WF) can provide a comprehensive knowledge of the use of water through the demarcation of the amount of the water consumed by different crops. The WF has three components: green (WFg), blue (WFb) and grey (WFgr) water footprints. The WFg refers to the rainwater stored in the root zone soil layer and is mainly utilized for agricultural, horticultural and forestry production. The WFb, however, is the consumptive use of water from surface or groundwater resources and mainly deals with irrigated agriculture, industry, domestic water use, etc. While the WFgr is the amount of fresh water required to assimilate pollutants resulting from the use of fertilizers/agrochemicals. This study was conducted on six agricultural fields in the Eastern region of Saudi Arabia, during the period from December 2015 to December 2016, to investigate the spatiotemporal variation of the WF of silage maize and carrot crops. The WF of each crop was estimated in two ways, namely agro-meteorological (WFAgro) and remote sensing (WFRS) methods. The blue, green and grey components of WFAgro were computed with the use of weather station/Eddy covariance measurements and field recorded crop yield datasets. The WFRS estimated by applying surface energy balance principles on Landsat-8 imageries. However, due to non-availability of Landsat-8 data on the event of rainy days, this study was limited to blue component (WFRS-b). The WFAgro of silage maize was found to range from 3545 m3 t-1 to 4960 m3 t-1; on an average, the WFAgro-g, WFAgro-b, and WFAgro-gr are composed of < 1%, 77%, and 22%, respectively. In the case of carrot, the WFAgro ranged between 297 m3 t-1 and 502 m3 t-1. The WFAgro-g of carrot crop was estimated at <1%, while WFAgro-b and WFAgro-gr was 67% and 32%, respectively. The WFAgro-b is occupied as a major portion in WF of silage maize (77%) and carrot (68%) crops. This is due to the high crop water demand combined with a very erratic rainfall, the irrigation is totally provided using groundwater delivered by center pivot irrigation systems. On the other hand, the WFRS-b estimated using Landsat-8 data was varied from 276 (±73) m3 t-1 (carrot) and 2885 (±441) m3 t-1 (silage maize). The variation (RMSE) between WFRS-b and WFAgro-b was about 17% and 14% for silage maize and carrot crops, respectively.

Estimation of gross primary production of irrigated maize using Landsat-8 imagery and Eddy Covariance data.

A study was conducted to understand the potential of Landsat-8 in the estimation of gross primary production (GPP) and to quantify the productivity of maize crop cultivated under hyper-arid conditions of Saudi Arabia. The GPP of maize crop was estimated by using the Vegetation Photosynthesis Model (VPM) utilizing remote sensing data from Landsat-8 reflectance (GPPVPM) as well as the meteorological data provided by Eddy Covariance (EC) system (GPPEC), for the period from August to November 2015. Results revealed that the cumulative GPPEC for the entire growth period of maize crop was 1871 g C m-2. However, the cumulative GPP determined as a function of the enhanced vegetation index - EVI (GPPEVI) was 1979 g C m-2, and that determined as a function of the normalized difference vegetation index - NDVI (GPPNDVI) was 1754 g C m-2. These results indicated that the GPPEVI was significantly higher than the GPPEC (R2 = 0.96, P = 0.0241 and RMSE = 12.6%). While, the GPPNDVI was significantly lower than the GPPEC (R2 = 0.93, P = 0.0384 and RMSE = 19.7%). However, the recorded relative error between the GPPEC and both the GPPEVI and the GPPNDVI was -6.22% and 5.76%, respectively. These results demonstrated the potential of the landsat-8 driven VPM model for the estimation of GPP, which is relevant to the productivity and carbon fluxes.

Prediction of Potato Crop Yield Using Precision Agriculture Techniques.

Crop growth and yield monitoring over agricultural fields is an essential procedure for food security and agricultural economic return prediction. The advances in remote sensing have enhanced the process of monitoring the development of agricultural crops and estimating their yields. Therefore, remote sensing and GIS techniques were employed, in this study, to predict potato tuber crop yield on three 30 ha center pivot irrigated fields in an agricultural scheme located in the Eastern Region of Saudi Arabia. Landsat-8 and Sentinel-2 satellite images were acquired during the potato growth stages and two vegetation indices (the normalized difference vegetation index (NDVI) and the soil adjusted vegetation index (SAVI)) were generated from the images. Vegetation index maps were developed and classified into zones based on vegetation health statements, where the stratified random sampling points were accordingly initiated. Potato yield samples were collected 2-3 days prior to the harvest time and were correlated to the adjacent NDVI and SAVI, where yield prediction algorithms were developed and used to generate prediction yield maps. Results of the study revealed that the difference between predicted yield values and actual ones (prediction error) ranged between 7.9 and 13.5% for Landsat-8 images and between 3.8 and 10.2% for Sentinel-2 images. The relationship between actual and predicted yield values produced R2 values ranging between 0.39 and 0.65 for Landsat-8 images and between 0.47 and 0.65 for Sentinel-2 images. Results of this study revealed a considerable variation in field productivity across the three fields, where high-yield areas produced an average yield of above 40 t ha-1; while, the low-yield areas produced, on the average, less than 21 t ha-1. Identifying such great variation in field productivity will assist farmers and decision makers in managing their practices.

Assessing the Spatial Variability of Alfalfa Yield Using Satellite Imagery and Ground-Based Data.

Understanding the temporal and spatial variability in a crop yield is viewed as one of the key steps in the implementation of precision agriculture practices. Therefore, a study on a center pivot irrigated 23.5 ha field in Saudi Arabia was conducted to assess the variability in alfalfa yield using Landsat-8 imagery and a hay yield monitor data. In addition, the study was designed to also explore the potential of predicting the alfalfa yield using vegetation indices. A calibrated yield monitor mounted on a large rectangular hay baler was used to measure the actual alfalfa yield for four alfalfa harvests performed in the period from October 2013 to May 2014. A total of 18 Landsat-8 images, representing different crop growth stages, were used to derive different vegetation indices (VIs). Data from the yield monitor was used to generate yield maps, which illustrated a definite spatial variation in alfalfa yield across the experimental field for the four studied harvests as indicated by the high spatial correlation values (0.75 to 0.97) and the low P-values (4.7E-103 to 8.9E-27). The yield monitor-measured alfalfa actual yield was compared to the predicted yield form the Vis. Results of the study showed that there was a correlation between actual and predicted yield. The highest correlations were observed between actual yield and the predicted using NIR reflectance, SAVI and NDVI with maximum correlation coefficients of 0.69, 0.68 and 0.63, respectively.