Exploitation of spatial information in high resolution digital imagery to map leaf area index

Crop responses to management practices and the environment, as quantified by leaf area index (LAI), provide decision-making criteria for the delineation of crop management zones. The objective of this work was to investigate whether spatial correlations inferred from remotely sensed imagery can be used to interpolate and map LAI using a relatively small number of ground-based LAI measurements. Airborne imagery was recorded with the Airborne Imaging Spectrometer for Applications (AISA) radiometer over a 3.2 ha corn field. Spectral vegetation indexes (SVI) were derived from the image and aggregated to cells of 2 × 2 m², 4 × 4 m², and 8 × 8 m² resolution. The residual maximum likelihood method was used to estimate the LAI variogram parameters. A generalized least squares regression was used to relate ground truth LAI data and collocated image pixels. This regression result was then used to convert variograms from the imagery to LAI units as well as to interpolate and map LAI. The decrease in resolution by merging pixels led to an increase in the value of the r ² and to a decrease in root mean-squared error (RMSE) values. The accuracy of kriged estimates from the variogram of the measured LAI and that from the image derived variograms was estimated by cross-validation. There was no difference in the accuracy of the estimates using either variograms from measured LAI values or from those of converted SVIs. Maps of LAI from ground-based measurements made by kriging the data with image-derived variogram parameters were similar to those obtained by with kriging with the variogram of measured LAI. Similar coarse spatial trends of high, medium and low LAI were evident for both maps. Variogram parameters from ground-based measurements of LAI compared favorably with those derived from remotely sensed imagery and could be used to provide reasonable results for the interpolation of LAI measurements.     

Investigating geostatistical methods to model within-field yield variability of cranberries for potential management zones

Cranberry harvesting methods give only one yield value per field making characterization of within-field variation, the usual first step in precision farming, difficult. Time-consuming berry count yield and fruit rot estimations are the best “ground truth” indication of yield variation within fields. Correlations and coincidence of binary classifications based on less expensive methods such as enhanced vegetation index (EVI) from imagery, and area to point (AtoP) kriging of useable, poor quality and trash yields were compared with this “ground truth”. In general AtoP kriged values gave higher correlations and kappa statistic values with berry counts and fruit rot than EVI. Geostatistical disaggregation of per field yield totals using AtoP kriging with EVI as an external drift (AtoP^KED) was also investigated. Factorial kriging was used to separate the several scales of variation in “ground truth” and EVI data and determine which ones were most spatially coherent/manageable and which related best to the AtoP kriged data. The spatial trend component of pre-harvest berry counts and AtoP kriging of yields both gave a good initial definition of spatially coherent, relatively permanent management zones. They were related to topography and depth of water table in the soil which are key factors governing cranberry yield. AtoP kriging or AtoP^KED are recommended for defining management zones as they are less expensive than berry counts. The value of AtoP kriging to precision farmers for other crops to map soils at the farm scale with some imagery and just one bulked soil sample per field or use nutrient levels associated with each polygon of traditional soil survey maps is discussed in the conclusions.     

Evaluation of methods to determine residual soil nitrate zones across the northern Great Plains of the USA

A four-year study was conducted from 2000 to 2004 at eight field sites in Montana, North Dakota and western Minnesota. Five of these sites were in North Dakota, two were in Montana and one was in Minnesota. The sites were diverse in their cropping systems. The objectives of the study were to (1) evaluate data from aerial photographs, satellite images, topographic maps, soil electrical conductivity (EC_a) sensors and several years of yield to delineate field zones to represent residual soil nitrate and (2) determine whether the use of data from several such sources or from a single source is better to delineate nitrogen management zones by a weighted method of classification. Despite differences in climate and cropping, there were similarities in the effectiveness of delineation tools for developing meaningful residual soil nitrate zones. Topographic information was usually weighted the most because it produced zones that were more correlated to actual soil residual nitrate than any other source of data at all locations. The soil EC_a sensor created better correlated zones at Minot, Williston and Oakes than at most eastern sites. Yield data for an individual year were sometimes useful, but a yield frequency map that combined several years of standardized yield data was more useful. Satellite imagery was better than aerial photographs at most locations. Topography, satellite imagery, yield frequency maps and soil EC_a are useful data for delineating nutrient management zones across the region. Use of two or more sources of data resulted in zones with a stronger correlation with soil nitrate.     

Airborne Videography to Identify Spatial Plant Growth Variability for Grain Sorghum

Much research has focused on the use of intensive grid soil sampling and yield monitors to identify within-field spatial variability in precision farming. This paper reports on the use of airborne videography to identify spatial plant growth patterns for grain sorghum. Color-infrared (CIR) digital video images were acquired from two grain sorghum fields in south Texas several times during the 1995 and 1996 growing seasons. The video images were registered, and classified into several zones of homogeneous spectral response using an unsupervised classification procedure. Ground truthing was performed upon a limited number of sites within each zone to determine plant density, plant height, leaf area index, biomass, and grain yield. Results from both years showed that the digital video imagery identified within-field plant growth variability and that classification maps effectively differentiated grain production levels and growth conditions within the two fields. A temporal comparison of the images and classification maps indicated that plant growth patterns differed somewhat between the two successive growing seasons, though areas exhibiting consistently high or low yield were identified within each field.     

Zone mapping application for precision-farming: a decision support tool for variable rate application

A web-based decision support tool, zone mapping application for precision farming (ZoneMAP, ), has been developed to automatically determine the optimal number of management zones and delineate them using satellite imagery and field data provided by users. Application rates, such as of fertilizer, can be prescribed for each zone and downloaded in a variety of formats to ensure compatibility with GPS-enabled farming equipment. ZoneMAP is linked to Digital Northern Great Plains, a web-based application which hosts an archive of satellite imagery, as well as high resolution imagery from airborne sensors. Management zones created by ZoneMAP mapped natural variation of the soil organic matter and other nutrients relatively well and were consistent with zone maps created by traditional means. The results demonstrated that ZoneMAP can serve as an effective and easy-to-use tool for those who practice precision agriculture.     

Site-specific management zones based on the Rasch model and geostatistical techniques

Delineation of management zones (MZ), i.e. areas within the field which represent subfield regions of similar production potential, is the first stage to implement site-specific management. During the last years different algorithms have been proposed to define MZ, with different results. In this work, the use of an objective method, the formulation of the Rasch model, which synthesizes data with different units into a uniform analytical framework, is considered to get representative measures of soil fertility potential which could be used to delimit MZ.To illustrate the method, a case study was conducted in a experimental field using five soil properties: clay, sand and silt content, and deep (ECd) and shallow (ECs) soil apparent electrical conductivity (approximately 0-90 and 0-30 cm depths, respectively). Two main results were obtained after applying this method: (1) a classification of all locations according to the soil fertility potential, which was the value of the Rasch measure and (2) the influence on the soil fertility of each individual soil property, being ECs the most influential and silt content the less influential property.Later, from the measures of soil fertility potential at sampled points, estimates were carried out using the ordinary kriging technique. Consequently, kriged estimates were utilized to map soil fertility potential and MZ were delimited using an equal-size classification method, which practically coincided with the MZ determined by a unsupervised classification.It is also shown the possibility of using probability maps to delimit MZ or provide information for hazard assessment of soil fertility in a field.     

Mobile TDR for geo-referenced measurement of soil water content and electrical conductivity

The development of site-specific crop management is constrained by the availability of sensors for monitoring important soil and crop related conditions. A mobile time-domain reflectometry (TDR) unit for geo-referenced soil measurements has been developed and used for detailed mapping of soil water content and electrical conductivity within two research fields. Measurements made during the early or late season, when soil moisture levels are close to field capacity, are related to the amount of plant available water and soil texture. Combined measurements of water content and electrical conductivity are closely related to the clay and silt fractions of a variable field. The application to early season field mapping of water content, electrical conductivity and clay content is presented. The water and clay content maps are to be used for automated delineation of field management units. Based on a spatial analysis of the soil water measurements, recommendations are made with respect to sampling strategies. Depending on the variability of a given area, between 15 and 30 ha can be mapped with respect to soil moisture and electrical conductivity with sufficient detail within 8 h.     

Frequency Analysis of Yield for Delineating Yield Response Zones

The yield in any given field or management zone is a product of interaction between many soil properties and production inputs. Therefore, multi-year yield maps may give better insight into determining potential management zones. This research was conducted to develop a methodology to delineate yield response zones by using two-state frequency analysis conducted on yield maps for 3 years on two commercial corn fields near Wiggins, Colorado. A zone was identified by the number of years that yield was equal and greater than the average yield in a given year. Classes producing statistically similar yield were combined resulting in three potential yield zones. Results indicated that the variability of yield over time and space could successfully be assessed at the same time without the drawbacks of averaging data from different years. Frequency analysis of multi-year yield data could be an effective way to establish yield response zones. Seventeen percent of the field #1 consistently produced lower yield than the mean while 43 of the field produced yield over the mean. Corresponding values for field #2 were 6% and 42%.The remainder of the fields produced fluctuating yields between years. These spatially and temporally sound yield response maps could be used to identify the yield-limiting factors in zones where yield is either low or fluctuating. Yield response maps could also be helpful to delineate potential management zones with the help of resource zones such as electrical conductivity and soil maps, along with the directed soil sampling results.     

Delineating site-specific management zones for pH-induced iron chlorosis

Iron chlorosis can limit crop yield, especially on calcareous soil. Typical management for iron chlorosis includes the use of iron fertilizers or chlorosis tolerant cultivars. Calcareous and non-calcareous soil can be interspersed within fields. If chlorosis-prone areas within fields can be predicted accurately, site-specific use of iron fertilizers and chlorosis-tolerant cultivars might be more profitable than uniform management. In this study, the use of vegetation indices (VI) derived from aerial imagery, on-the-go measurement of soil pH and apparent soil electrical conductivity (EC_a) were evaluated for their potential to delineate chlorosis management zones. The study was conducted at six sites in 2004 and 2005. There was a significant statistical relationship between grain yield and selected properties at two sites (sites 1 (2005) and 3), moderate relationships at sites 2 and 4, and weak relationships at site 5. For sites 1 (2005) and 3, and generally across all sites, yield was predicted best with the combination of NDVI and deep EC_a. These two properties were used to delineate chlorosis management zones for all sites. Sites 1 and 3 showed a good relationship between delineated zones and the selected properties, and would be good candidates for site-specific chlorosis management. For site 5, differences in the properties between mapped zones were small, and the zones had weak relationships to yield. This site would be a poor candidate for site-specific chlorosis management. Based on this study, the delineation of chlorosis management zones from aerial imagery combined with soil EC_a appears to be a useful tool for the site-specific management of iron chlorosis.     

Mapping radiation interception in row-structured orchards using 3D simulation and high-resolution airborne imagery acquired from a UAV

This study was conducted to model the fraction of intercepted photosynthetically active radiation (fIPAR) in heterogeneous row-structured orchards, and to develop methodologies for accurate mapping of the instantaneous fIPAR at field scale using remote sensing imagery. The generation of high-resolution maps delineating the spatial variation of the radiation interception is critical for precision agriculture purposes such as adjusting management actions and harvesting in homogeneous within-field areas. Scaling-up and model inversion methods were investigated to estimate fIPAR using the 3D radiative transfer model, Forest Light Interaction Model (FLIGHT). The model was tested against airborne and field measurements of canopy reflectance and fIPAR acquired on two commercial peach and citrus orchards, where study plots showing a gradient in the canopy structure were selected. High-resolution airborne multi-spectral imagery was acquired at 10?nm bandwidth and 150?mm spatial resolution using a miniaturized multi-spectral camera on board an unmanned aerial vehicle (UAV). In addition, simulations of the land surface bidirectional reflectance were conducted to understand the relationships between canopy architecture and fIPAR. Input parameters used for the canopy model, such as the leaf and soil optical properties, canopy architecture, and sun geometry were studied in order to assess the effect of these inputs on canopy reflectance, vegetation indices and fIPAR. The 3D canopy model approach used to simulate the discontinuous row-tree canopies yielded spectral RMSE values below 0.03 (visible region) and below 0.05 (near-infrared) when compared against airborne canopy reflectance imagery acquired over the sites under study. The FLIGHT model assessment conducted for fIPAR estimation against field measurements yielded RMSE values below 0.08. The simulations conducted suggested the usefulness of these modeling methods in heterogeneous row-structured orchards, and the high sensitivity of the normalized difference vegetation index and fIPAR to background, row orientation, percentage cover and sun geometry. Mapping fIPAR from high-resolution airborne imagery through scaling-up and model inversion methods conducted with the 3D model yielded RMSE error values below 0.09 for the scaling-up approach, and below 0.10 for the model inversion conducted with a look-up table. The generation of intercepted radiation maps in row-structured tree orchards is demonstrated to be feasible using a miniaturized multi-spectral camera on board UAV platforms for precision agriculture purposes.     

Evaluating unoccupied aerial systems (UAS) imagery as an alternative tool towards cotton-based management zones

Unoccupied aerial system (UAS) imagery may serve as an additional tool towards management zone delineation. This is because UAS data collection is relatively flexible. However, it is unclear how useful UASs can be towards generating management zones, relative to preexisting tools (e.g. apparent soil electrical conductivity or EC_a). The purpose of this study, therefore, was to evaluate UAS imagery, relative to EC_a, in terms of their ability to: 1) predict cotton traits (i.e. height, seed cotton yield), and 2) define cotton management zones based on these traits. Single-season UAS images from multispectral/thermal sensors were collected and processed into Normalized Difference Vegetation Index (NDVI) and radiometric surface temperature (T_r), respectively. Management zones were also delineated using digital camera (RGB) imagery collected at periods before planting and near harvest. RGB management zones were delineated by a novel open boll mapping approach. In-season NDVI and T_r layers were significant (P?<?0.01) predictors of canopy height. Additionally, NDVI and T_r maps produced statistically different management zones during flowering and boll filling growth stages in terms of yield (P?=?0.001 or less). Open boll layers were all more accurate predictors of cotton seed yield than EC_a data—these two layers also produced statistically distinct management zones. ANOVA tests revealed that, given EC_a alone, adding UAS information via the RGB open boll map resulted in a significantly different yield prediction model (P?<?0.001). These results suggest that UAS imagery can offer valuable information for cotton management zone delineation that other techniques cannot.

     

Prediction of leaf area index in almonds by vegetation indexes

Three levels of scale for determining leaf area index (LAI) were explored within an almond orchard of alternating rows of Nonpareil and Monterey varieties using hemispherical photography and mule lightbar (MLB) at ground level up to airborne and satellite imagery. We compared LAI estimates of 56 fisheye photos strategically placed in the orchard to validate 500,000 MLB point scans of a small portion of the aisles between tree rows to water and vegetation indexes of MASTER (MODIS/ASTER simulator) and Landsat 5 imagery. The high correlation of fisheye photo LAI to MLB LAI estimates establishes this new method against the measurement standard within the plant community while significantly increasing sample size. MLB LAI and MASTER vegetation indexes, such as NDWI (normalized difference water index), GMI (Gitelson–Merzlyak index) and NDVI (normalized difference vegetation index), were highly correlated (r² = 0.90). In addition, a high correlation (r² = 0.80) between the MLB measured LAI and selected Landsat derived vegetation indexes (VI) was found. This scaling and validation of LAI estimate expands the spatial area and frequency of determination for time series analysis of crop phenology studies.     

Impact of soil types on sugarcane development monitored over time by remote sensing

Soil is one of the most important factors for agricultural production. In tropical regions, soil variability is considerable, with the most diverse combinations of physical and chemical characteristics, an influence factor in crop growth and productivity. In this research, the main objective was to identify how soil characteristics and parent material can influence sugarcane development over time using remote sensing. An area located in Sao Paulo, Brazil, of 182 ha (one point per ha with soil analysis), with high variability in the parent material and soil types, was selected. Images from the Sentinel2-MSI satellite were used to describe the spectral behavior of sugarcane over a period of one year. The NDRE (normalized difference red-edge index) was calculated for each image and then the leaf area index (LAI) was obtained from it. Maps of soil classes, soil properties at two depths (0–0.20 and 0.80–1.0 m), and parent material classes were related to sugarcane LAI variability over time. Production environment zones, which is a classification based on soil characteristics to support sugarcane development, were also obtained and related to LAI variability. Spectral signatures of the crop presented different behaviors through the season, soil types and soil attributes provided useful responses for this variability. At the beginning of the season, the surface and subsurface soil properties (texture and fertility) impacted differently on crop development. On the other hand, soil classes and parent material influenced LAI in all production environments studied. The results indicated that the soil types and their properties at different depths have a significant impact on sugarcane development. Furthermore, RS was able to monitor the plant evolution and be related to soil types which may assist in plant management. The results can bring light on how better sugarcane management can be conducted using remote sensing data and soils variability.

     

Field methods for making productivity classes for site-specific management of wheat

Reducing the decision-making unit to classes within fields can improve yields, efficiency in the use of nutrients and profitability of crops. The objectives were to compare methods for class delimitation in wheat (Triticum aestivum L.) crops based on apparent productivity levels and establish similarities among them in terms of spatial overlapping, productive attributes and the use of nitrogen. In three wheat fields, high and low apparent productivity classes (APC) were defined based on eight methodologies: yield maps, soil maps, gramineae vegetation index, rotation crop index, interpretation of satellite images, management records, elevation and integrated soil and yield maps. In each APC, soil and crop yield components were determined under five nitrogen fertilization levels. Among delimitation methodologies, the degree of coincidence varied from 1.4 to 81.7%. The differences in soil properties, nitrogen use efficiency and grain yields were greater among fields than among APC within each field. In each field, the delimitation methodologies identified different single factors that discriminated among the potential management classes and were partially associated with the crop grain yields. The wheat crops at the low APC yielded 39% less and 12% less than at the high APC, respectively. The nitrogen fertilization, at the rate for maximum productivity for each ACP, reduced the yield differences between contrasting APC. Nitrogen fertilization also modified clustering of classes based on expected yields. Making management classes for wheat based on expected productivity is more accurate when based on previous crop production information under similar nitrogen fertilization conditions than the targeted crop.

     

Ground-Sensor Soil Reflectance as Related to Soil Properties and Crop Response in a Cotton Field

Bare soil reflectance from airborne imagery or laboratory spectrometers has been used to infer soil properties such as soil texture, organic matter, water content, salinity and crop residue cover. However, the relation of soil properties to reflectance data often varies with soil type and conditions and surface reflectance may not be representative of the conditions in the root zone. The objectives of this study were to assess the soil reflectance data obtained by ground-based sensors and to model soil properties in the root zone as a function of surface soil reflectance and plant response. Ground-based sensors were used to simultaneously monitor soil and canopy reflectance in the visible and near-infrared (VNIR) along six rows and in two growth stages in a 7 ha cotton field. The reflectance data were compared to soil properties, leaf nutrients and biomass measured at 33 sampling positions along the rows. Brightness values of the blue and green bands of soil reflectance were better correlated to soil water content, particulate organic matter and extractable potassium and phosphorus, while those in the red and NIR bands were correlated to soil carbonate content, total nitrogen, electrical conductivity and foliar nutrients. The correlation of red soil reflectance with canopy reflectance was significant and indicated an indirect inverse relationship between soil fertility and plant stress. The integration of surface soil reflectance and plant response variables in a multiple regression model did not substantially improve the prediction of soil properties in the root zone. However, crop nutrient status explained a significant portion of the spatial variability of soil properties related to nitrification processes when soil reflectance did not. The implication of these findings to agricultural management is discussed.     

Management zones delineation using fuzzy clustering techniques in grapevines

Precision viticulture aims at managing vineyards at a sub-field scale according to the real needs of each part of the field. The current study focused on delineating management zones using fuzzy clustering techniques and developing a simplified approach for the comparison of zone maps. The study was carried out in a 1.0 ha commercial vineyard in Central Greece during 2009 and 2010. Variation of soil properties across the field was initially measured by means of electrical conductivity, soil depth and topography. To estimate grapevine canopy properties, NDVI was measured at different stages during the vine growth cycle. Yield and grape composition (must sugar content and total acidity) mapping was carried out at harvest. Soil properties, yield and grape composition parameters showed high spatial variability. All measured data were transformed on a 48-cell grid (10 × 20 m) and maps of two management zones were produced using the MZA software. Pixel-by-pixel comparison between maps of electrical conductivity, elevation, slope, soil depth and NDVI with yield and grape composition maps, set as reference parameters, allowed for the calculation of the degree of agreement, i.e. the percentage of pixels belonging to the same zone. The degree of agreement was used to select the best-suited parameters for final management zones delineation. For the year 2009 soil depth, early and mid season NDVI were used for yield-based management zones while for quality-based management zones ECa, early and mid season NDVI were utilized. For the year 2010 ECa, elevation and NDVI acquired during flowering and veraison were used for the delineation of yield-based management zones while for quality-based management zones ECa and NDVI acquired during flowering and harvest were utilized. Results presented here could be the basis for simple management zone delineation and subsequent improved vineyard management.     

Sensing technologies for precision specialty crop production

With the advances in electronic and information technologies, various sensing systems have been developed for specialty crop production around the world. Accurate information concerning the spatial variability within fields is very important for precision farming of specialty crops. However, this variability is affected by a variety of factors, including crop yield, soil properties and nutrients, crop nutrients, crop canopy volume and biomass, water content, and pest conditions (disease, weeds, and insects). These factors can be measured using diverse types of sensors and instruments such as field-based electronic sensors, spectroradiometers, machine vision, airborne multispectral and hyperspectral remote sensing, satellite imagery, thermal imaging, RFID, and machine olfaction system, among others. Sensing techniques for crop biomass detection, weed detection, soil properties and nutrients are most advanced and can provide the data required for site specific management. On the other hand, sensing techniques for diseases detection and characterization, as well as crop water status, are based on more complex interaction between plant and sensor, making them more difficult to implement in the field scale and more complex to interpret. This paper presents a review of these sensing technologies and discusses how they are used for precision agriculture and crop management, especially for specialty crops. Some of the challenges and considerations on the use of these sensors and technologies for specialty crop production are also discussed.     

QuickBird satellite versus ground-based multi-spectral data for estimating nitrogen status of irrigated maize

In-season nitrogen (N) management of irrigated maize (Zea mays L.) requires frequent acquisition of plant N status estimates to timely assess the onset of crop N deficiency and its spatial variability within a field. This study compared ground-based Exotech nadir-view sensor data and QuickBird satellite multi-spectral data to evaluate several green waveband vegetation indices to assess the N status of irrigated maize. It also sought to determine if QuickBird multi-spectral imagery could be used to develop plant N status maps as accurately as those produced by ground-based sensor systems. The green normalized difference vegetation index normalized to a reference area (NGNDVI) clustered the data for three clear-day data acquisitions between QuickBird and Exotech data producing slopes and intercepts statistically not different from 1 and 0, respectively, for the individual days as well as for the combined data. Comparisons of NGNDVI and the N Sufficiency Index produced good correlation coefficients that ranged from 0.91 to 0.95 for the V12 and V15 maize growth stages and their combined data. Nitrogen sufficiency maps based on the NGNDVI to indicate N sufficient (≥0.96) or N deficient (<0.96) maize were similar for the two sensor systems. A quantitative assessment of these N sufficiency maps for the V10–V15 crop growth stages ranged from 79 to 83% similarity based on areal agreement and moderate to substantial agreement based on the kappa statistics. Results from our study indicate that QuickBird satellite multi-spectral data can be used to assess irrigated maize N status at the V12 and later growth stages and its variability within a field for in-season N management. The NGNDVI compensated for large off-nadir and changing target azimuth view angles associated with frequent QuickBird acquisitions.