Murilo M. Maeda

Cotton growth modeling and assessment using unmanned aircraft system visual-band imagery

Abstract. This paper explores the potential of using unmanned aircraft system (UAS)-based visible-band images to assess cotton growth. By applying the structure-from-motion algorithm, the cotton plant height (ph) and canopy cover (cc) information were retrieved from the point cloud-based digital surface models (DSMs) and orthomosaic images. Both UAS-based ph and cc follow a sigmoid growth pattern as confirmed by ground-based studies. By applying an empirical model that converts the cotton ph to cc, the estimated cc shows strong correlation (R2=0.990) with the observed cc. An attempt for modeling cotton yield was carried out using the ph and cc information obtained on June 26, 2015, the date when sigmoid growth curves for both ph and cc tended to decline in slope. In a cross-validation test, the correlation between the ground-measured yield and the estimated equivalent derived from the ph and/or cc was compared. Generally, combining ph and cc, the performance of the yield estimation is most comparable against the observed yield. On the other hand, the observed yield and cc-based estimation produce the second strongest correlation, regardless of the complexity of the models.

Unmanned aircraft system-derived crop height and normalized difference vegetation index metrics for sorghum yield and aphid stress assessment

Abstract. A small, fixed-wing unmanned aircraft system (UAS) was used to survey a replicated small plot field experiment designed to estimate sorghum damage caused by an invasive aphid. Plant stress varied among 40 plots through manipulation of aphid densities. Equipped with a consumer-grade near-infrared camera, the UAS was flown on a recurring basis over the growing season. The raw imagery was processed using structure-from-motion to generate normalized difference vegetation index (NDVI) maps of the fields and three-dimensional point clouds. NDVI and plant height metrics were averaged on a per plot basis and evaluated for their ability to identify aphid-induced plant stress. Experimental soil signal filtering was performed on both metrics, and a method filtering low near-infrared values before NDVI calculation was found to be the most effective. UAS NDVI was compared with NDVI from sensors onboard a manned aircraft and a tractor. The correlation results showed dependence on the growth stage. Plot averages of NDVI and canopy height values were compared with per-plot yield at 14% moisture and aphid density. The UAS measures of plant height and NDVI were correlated to plot averages of yield and insect density. Negative correlations between aphid density and NDVI were seen near the end of the season in the most damaged crops.

Temporal Estimates of Crop Growth in Sorghum and Maize Breeding Enabled by Unmanned Aerial Systems

We comprehensively validated the use of UAS in sorghum and maize breeding programs. Temporal estimates of plant growth will allow researchers to elucidate new phenotypes. The stage of the breeding pipeline dictates the applicability of UAS platforms. The implementation of UAS is demonstrated in different crop species. Monetary and time costs should be considered before implementation of UAS.

A ground based platform for high throughput phenotyping

Abstract The objective of this effort was to evaluate current commercially-available sensor technology (three sonic ranging and two NDVI sensors) for use in a ground-based platform for plant phenotyping and crop management decisions. The Global Positioning System (GPS) receiver from Trimble provided a high level of accuracy during our tests. Normalized Difference Vegetation Index (NDVI) data collected using the GreenSeeker sensors were more consistent and presented less variability when compared to the Decagon SRS sensor. The consistency could be due to the GreenSeeker system averaging readings of more rows. The tests also indicated that although sonic ranging sensor technology may be employed to obtain average plant height estimates, the technology is still a limiting factor for high-accuracy measurements at the plant level.

Unmanned aerial system assisted framework for the selection of high yielding cotton genotypes

Abstract Recent advances in molecular breeding and bioinformatics have greatly accelerated the screening of large sets of genotypes. Development of field-level phenotyping, however, still lags behind and is considered by many as the main bottleneck to improved efficiency in breeding programs. Unmanned Aerial System (UAS) and sensor technology available today enables collection of data at high spatial and temporal scales, previously unobtainable using traditional airborne remote sensing technologies. Here, we propose an UAS-assisted high throughput phenotyping framework for cotton ( Gossypium hirsutum L.) genotype selection. UAS data collected on July 24, 2015 were used to calculate canopy cover, and UAS data collected on August 5, 2015 were used to extract open boll related phenotypic features including number of open bolls, average area of open bolls, average diameter of open bolls, perimeter of open bolls, perimeter to area ratio. Using the extracted features, a sequential selection procedure was performed on a population of 144 entries. Entries selected from the proposed framework were compared to the highest yielding entries determined by mechanical harvest results. Experimental results indicated that the selection process increased minimum and average lint yield of the remaining population by 7.4 and 10%, respectively, and UAS-selected entries and genotypes matched 80 and 73%, respectively, the same lists ranked by actual field harvest measurements.

Phenotyping of sorghum panicles using unmanned aerial system (UAS) data

Unmanned Aerial System (UAS) is getting to be the most important technique in recent days for precision agriculture and High Throughput Phenotyping (HTP). Attributes of sorghum panicle, especially, are critical information to assess overall crop condition, irrigation, and yield estimation. In this study, it is proposed a method to extract phenotypes of sorghum panicles using UAS data. UAS data were acquired with 85% overlap at an altitude of 10m above ground to generate super high resolution data. Orthomosaic, Digital Surface Model (DSM), and 3D point cloud were generated by applying the Structure from Motion (SfM) algorithm to the imagery from UAS. Sorghum panicles were identified from orthomosaic and DSM by using color ratio and circle fitting. The cylinder fitting method and disk tacking method were proposed to estimate panicle volume. Yield prediction models were generated between field-measured yield data and UAS-measured attributes of sorghum panicles.

Cotton growth modeling using unmanned aerial vehicle vegetation indices

Unmanned aerial vehicle (UAV) images have great potential for agricultural researches because of their high spatial and temporal resolutions. However, most UAV researches in the agriculture field have adopted vegetation indices without second derivative parameters related with a growth model. In addition, visible band vegetation indices in UAV researches have not been explored in detail despite of their importance in UAV application. In this study, three RGB vegetation indices that showed good performance in previous studies are adopted and growth modeling using time series vegetation indices is proposed. In addition, growth model-based second derivatives are extracted for crop growth analysis. R squares of the proposed method from three RGB vegetation indices were 0.8–0.9 and excessive green vegetation index (ExG) showed the best accuracy.

Sorghum panicle extraction from unmanned aerial system data

Attributes of sorghum panicle is a very important to assess overall crop condition, irrigation, and estimation of terminal yield. In this study, a novel method to extract sorghum panicle and estimate panicle volume of grain sorghum using an Unmanned Aerial System (UAS) is proposed. UAS data were acquired with 85% overlap at an altitude of 10m above ground. Ortho-mosaic image, Digital Surface Model (DSM), and 3D point cloud were generated by applying the Structure from Motion (SfM) algorithm to the images. Ground Control Points (GCP) were used for accurate geo-referencing. Sorghum panicles were determined from RGB image and DSM by using color ratio and circle fitting. Panicle volume was estimated by the cylinder fitting method and the disk stacking method. The results of this study showed that UAS data can provide non-destructive, more efficient, and may be considered to replace the field work.