Browsing by Author "Yule IJ"

Now showing 1 - 9 of 9
  • Thumbnail Image
    Item
    A physically informed multi-scale deep neural network for estimating foliar nitrogen concentration in vegetation
    (Elsevier B.V., 2024-05-28) Dehghan-Shoar MH; Kereszturi G; Pullanagari RR; Orsi AA; Yule IJ; Hanly J
    This study introduces a Physically Informed Deep Neural Network (PINN) that leverages spectral data and Radiative Transfer Model insights to improve nitrogen concentration estimation in vegetation, addressing the complexities of physical processes. Utilizing a comprehensive spectroscopy dataset from various species across dry/ground (n = 2010), leaf (n = 1512), and canopy (n = 6007) scales, the study identifies 13 spectral bands key for chlorophyll and protein quantification. Key bands at 2276 nm, 755 nm, 1526 nm, 2243 nm, and 734 nm emerged vital for accurate N% prediction. The PINN outperforms partial least squares regression and standard deep neural networks, achieving an R2 of 0.71 and an RMSE of 0.42 (%N) on an independent validation set. Results indicate dry/ground data performed best (R2 = 0.9, RMSE = 0.24 %N), with leaf and canopy data showing lower efficacy (R2 = 0.67, RMSE = 0.45 %N; R2 = 0.65, RMSE = 0.46 %N, respectively). This multi-scale approach provides insights into spectral and N% relationships, enabling precise estimation across vegetation types and facilitating the development of transferable models. The PINN offers a new avenue for analyzing remote sensing data, demonstrating the significant potential for accurate, scale-spanning N% estimation in vegetation. Further enriching our analysis, the inclusion of seasonal data significantly enhanced our model's performance in field spectroscopy, with notable improvements observed across summer, spring, autumn, and winter. This adjustment underlines the model's increased accuracy and predictive capability at the field spectroscopy scale, emphasizing the vital role of integrating environmental factors, including climatic and physiological aspects, in future research.
  • Thumbnail Image
    Item
    A Unified Physically Based Method for Monitoring Grassland Nitrogen Concentration with Landsat 7, Landsat 8, and Sentinel-2 Satellite Data
    (MDPI (Basel, Switzerland), 2023-05-09) Dehghan-Shoar MH; Pullanagari RR; Kereszturi G; Orsi AA; Yule IJ; Hanly J; Berger K; Croft H; Liu T; Lu B; Yin D
    The increasing number of satellite missions provides vast opportunities for continuous vegetation monitoring, crucial for precision agriculture and environmental sustainability. However, accurately estimating vegetation traits, such as nitrogen concentration (N%), from Landsat 7 (L7), Landsat 8 (L8), and Sentinel-2 (S2) satellite data is challenging due to the diverse sensor configurations and complex atmospheric interactions. To address these limitations, we developed a unified and physically based method that combines a soil–plant–atmosphere radiative transfer (SPART) model with the bottom-of-atmosphere (BOA) spectral bidirectional reflectance distribution function. This approach enables us to assess the effect of rugged terrain, viewing angles, and illumination geometry on the spectral reflectance of multiple sensors. Our methodology involves inverting radiative transfer model variables using numerical optimization to estimate N% and creating a hybrid model. We used Gaussian process regression (GPR) to incorporate the inverted variables into the hybrid model for N% prediction, resulting in a unified approach for N% estimation across different sensors. Our model shows a validation accuracy of 0.35 (RMSE %N), a mean prediction interval width (MPIW) of 0.35, and an R (Formula presented.) of 0.50, using independent data from multiple sensors collected between 2016 and 2019. Our unified method provides a promising solution for estimating N% in vegetation from L7, L8, and S2 satellite data, overcoming the limitations posed by diverse sensor configurations and complex atmospheric interactions.
  • Thumbnail Image
    Item
    Ballistic modeling and pattern testing to prevent separation of New Zealand fertilizer products
    (American Society of Agricultural and Biological Engineers, 18/06/2015) Grafton MCE; Yule IJ; Robertson BG; Chok SM; Manning MJ
    In recent years twin disc centrifugal spreaders have become larger with some manufacturers claiming to be able to spread fertilizer products as far as 60 m. To achieve wider spread widths, the fertilizer particle exit velocity off the disc has increased, as a result the ballistic qualities of the product becomes more critical. This case study uses data-mined information from Ravensdown Fertiliser Co-op Ltd, a major fertilizer supplier. This article examines and researches products used by arable and grassland farmers and studies the effect of changes in product characteristics on spread bout width from these newer spreaders. Ballistic modeling, based on particle density, size, and shape was used to test the distance fertilizer particles travel at various velocities. Fertilizer particle velocities were measured by high speed photometry using both common fertilizers and common spreaders found in New Zealand. Spreading equipment was pattern tested using the New Zealand Spreadmark method. Ballistic modeling of particles proved appropriate in ideal conditions. Fertilizer manufacturers believe that spreader operators often fail to take account of physical characteristics of products being spread and target the widest bout width possible. This can lead to an in-field Coefficient of Variation (CV) which is much greater than 15% and leads to sub-optimal utilization of fertilizer, where variations in particle size distribution occur. Similar situations have been experienced when spreading fertilizer blends; where blends previously spread successfully, at narrower bout widths now separate. Ballistic models could provide bout width recommendations for products and blends, for a range of applicators and reduce crop striping.
  • Thumbnail Image
    Item
    Capability of ground fertiliser placement when spread from a fixed wing aircraft
    (13/04/2016) Chok S; Grafton M; Yule IJ; Manning M
    Aerial topdressing using differential rate application technology improves fertiliser spreading on hill country farms. However, the system’s ability to place fertiliser accurately and precisely within an area needs to be determined. Accuracy was determined by comparing measured and intended application rates. Precision was indicated by the coefficient of variation (CV), which is the standard deviation of the measured application rate over the mean of this rate. Two trials were carried out, where aircraft deposited fertiliser at two application rates and fertiliser was captured using cone-shaped collectors. The average measured application rate for both trials was less than the intended rate. The CV ranged from 35 to 57%, and was lower than CV’s from pilot-operated hopper systems (78%). A one-way analysis of variance test found the difference between measured application rate in the high and low application zone was statistically significant. The results indicate work is required to improve the accuracy and precision of the differential rate system, however, the system shows promise. Keywords: differential rate application technology, aerial spreading, fertiliser placement
  • Thumbnail Image
    Item
    Identifying grass species using hyperspectral sensing
    (Fertilizer and Lime Research Centre, ) Cushnahan T; Yule IJ; Pullanagari RR; Grafton MCE; Currie, L; Singh, R
  • Thumbnail Image
    Item
    Mapping nutrient concentration in pasture using hyperspectral imaging
    (2015) Yule IJ; Pullanagari RR; Irwin M; McVeagh P; Kereszturi G; White M; Manning M
  • Thumbnail Image
    Item
    Measuring the spread patterns of spreaders under normal field conditions compared to test conditions
    (Fertilizer and Lime Research Centre, 1/06/2016) Grafton MCE; Acebes DI; Yule IJ; Willis LA; Currie, L; Singh, R
  • Thumbnail Image
    Item
    The ballistics of separation of fertiliser blends at wide bout widths
    (28/02/2014) Grafton MCE; Yule IJ; Currie, LD; Christensen, CL
  • Thumbnail Image
    Item
    Understanding soil phosphorus variability with depth for the improvement of current soil sampling methods
    (7/02/2017) Kaul TM; Grafton MCE; Hedley MJ; Yule IJ
    Noise in soil test results can be reduced by measuring phosphorus below the top 3cm of soil from ground level. This is significant for improving current soil nutrient testing methods by allowing better geospatial predictions for whole paddock soil nutrient variability mapping for use in precision fertilizer application. In this study 200 cores were collected from predetermined grids at two trial sites at „Patitapu‟ hill country farm in the Wairarapa. The sites were selected according to accessibility and slope- Trial 1 was a 200m x 100m grid located in a gently undulating paddock. Trial 2 was a 220m x 80m grid located on a moderate to steeply sloped paddock. Each grid had cores taken at intervals of 5m, 10m and 20m. Core sites were mapped out on a Landsat 8 image (NASA) of the Trial sites using ArcGIS 10.2 (ESRI, Redlands Ca.) prior to going into the field; these were then marked out using a LEICA (real time kinematic GPS), pigtails and spray-paint on the ground. Cores were taken using a 30mm diameter soil core sampler. Trial 1 cores were cut into four sections according to depth: A – 0-30mm, B – 30mm-75mm, C- 75mm-150mm, and D- >150mm. Trial 2 cores were cut into three sections: A – 0-30mm, B – 30mm-75mm, C- 75mm-150mm. Olsen P lab results were collected for 120 of the 400 soil cores. These results were analyzed to compare the spatial variability of each depth. The results indicate that there is a significant decrease in variability from section A to section B for both trials. Section B and C for trial 1 have similar variability, whereas there is another significant drop in variability from section B to C in trial 2. Measuring samples below the top 3cm appears to effectively reduce noise, however measuring below 7.5cm for a steeply sloped paddock such as trial 2 may reduce variability too much as to no longer be representative of plant available P, and therefore misrepresenting the overall variability of soil P across a paddock or farm.