Search

Maturity is an important trait in dry pea breeding programs, but the conventional process predominately used to measure this trait can be time-consuming, labor-intensive, and prone to errors. Therefore, a more efficient and accurate approach would be desirable to support dry pea breeding programs. This study presents a novel approach for measuring dry pea maturity using machine learning algorithms and unmanned aerial systems (UASs)-collected data. We evaluated the abilities of five machine learning algorithms (random forest, artificial neural network, support vector machine, K-nearest neighbor, and naïve Bayes) to accurately predict dry pea maturity on field plots. The machine learning algorithms considered a range of variables, including crop height metrics, narrow spectral bands, and 18 distinct color and spectral vegetation indices. Backward feature elimination was used to select the most important features by iteratively removing insignificant ones until the model's predictive performance was optimized. The study's findings reveal that the most effective approach for assessing dry pea maturity involved a combination of narrow spectral bands, red-edge, near-infrared (NIR), and RGB-based vegetation indices, along with image textural metrics and crop height metrics. The implementation of a random forest model further enhanced the accuracy of the results, exhibiting the highest level of accuracy with a 0.99 value for all three metrics precision, recall, and f1 scores. The sensitivity analysis revealed that spectral features outperformed structural features when predicting pea maturity. While multispectral cameras achieved the highest accuracy, the use of RGB cameras may still result in relatively high accuracy, making them a practical option for use in scenarios where cost is a limiting factor. In summary, this study demonstrated the effectiveness of coupling machine learning algorithms, UASs-borne LIDAR, and multispectral data to accurately assess maturity in peas. [ABSTRACT FROM AUTHOR]

Forest disturbance detection is of great significance for understanding forest dynamics. The Landsat-based detection of the Trends in Disturbance and Recovery (LandTrendr) algorithm is widely used for forest disturbance mapping. However, there are still two limitations in LandTrendr: first, it only used for summer-composited observations, which may delay the detection of forest disturbances that occurred in autumn and winter by one year, and second, it detected all disturbance types simultaneously using a single spectral index, which may reduce the mapping accuracy for certain forest disturbance types. Here, we modified LandTrendr (mLandTrendr) for forest disturbance mapping in China by using multi-season observations and multispectral indices. Validations using the randomly selected 1957 reference forest disturbance samples across China showed that the overall accuracy (F1 score) of forest disturbance detection in China was improved by 21% with these two modifications. The mLandTrendr can quickly and accurately detect forest disturbance and can be extended to national and global forest disturbance mapping for various forest types. [ABSTRACT FROM AUTHOR]

The use of multispectral satellite imagery for water monitoring is a fast and cost-effective method that can benefit from the growing availability of medium–high-resolution and free remote sensing data. Since the 1970s, multispectral satellite imagery has been exploited by adopting different techniques and spectral indices. The high number of available sensors and their differences in spectral and spatial characteristics led to a proliferation of outcomes that depicts a nice picture of the potential and limitations of each. This paper provides a review of satellite remote sensing applications for water extent delineation and flood monitoring, highlighting trends in research studies that adopted freely available optical imagery. The performances of the most common spectral indices for water segmentation are qualitatively analyzed and assessed according to different land cover types to provide guidance for targeted applications in specific contexts. The comparison is carried out by collecting evidence obtained from several applications identifying the overall accuracy (OA) obtained with each specific configuration. In addition, common issues faced when dealing with optical imagery are discussed, together with opportunities offered by new-generation passive satellites. [ABSTRACT FROM AUTHOR]

The objective of this paper was to evaluate the accuracy of two advanced blending algorithms, Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) and Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model (ESTARFM) to downscale Moderate Resolution Imaging Spectroradiometer (MODIS) indices to the spatial resolution of Landsat. We tested two approaches: (i) “Index-then-Blend” (IB); and (ii) “Blend-then-Index” (BI) when simulating nine indices, which are widely used for vegetation studies, environmental moisture assessment and standing water identification. Landsat-like indices, generated using both IB and BI, were simulated on 45 dates in total from three sites. The outputs were then compared with indices calculated from observed Landsat data and pixel-to-pixel accuracy of each simulation was assessed by calculating the: (i) bias; (ii) R2; and (iii) Root Mean Square Deviation (RMSD). The IB approach produced higher accuracies than the BI approach for both blending algorithms for all nine indices at all three sites. We also found that the relative performance of the STARFM and ESTARFM algorithms depended on the spatial and temporal variances of the Landsat-MODIS input indices. Our study suggests that the IB approach should be implemented for blending of environmental indices, as it was: (i) less computationally expensive due to blending single indices rather than multiple bands; (ii) more accurate due to less error propagation; and (iii) less sensitive to the choice of algorithm.

The Biosphere Reserve of La Mancha Húmeda is a wetland-rich area located in central Spain. This reserve comprises a set of temporary lakes, often saline, where water level fluctuates seasonally. Water inflows come mainly from direct precipitation and runoff of small lake watersheds. Most of these lakes lack surface outlets and behave as endorheic systems, where water withdrawal is mainly due to evaporation, causing salt accumulation in the lake beds. Remote sensing was used to estimate the temporal variation of the flooded area in these lakes and their associated hydrological patterns related to the seasonality of precipitation and evapotranspiration. Landsat 7 ETM+ satellite images for the reference period 2013-2015 were jointly used with ground-truth datasets. Several inverse modeling methods, such as two-band and multispectral indices, single-band threshold, classification methods, artificial neural network, support vector machine and genetic programming, were applied to retrieve information on the variation of the flooded areas. Results were compared to ground-truth data, and the classification errors were evaluated by means of the kappa coefficient. Comparative analyses demonstrated that the genetic programming approach yielded the best results, with a kappa value of 0.98 and a total error of omission-commission of 2%. The dependence of the variations in the water-covered area on precipitation and evaporation was also investigated. The results show the potential of the tested techniques to monitor the hydrological patterns of temporary lakes in semiarid areas, which might be useful for management strategy-linked lake conservation and specifically to accomplish the goals of both the EuropeanWater Framework Directive and the Habitats Directive. [ABSTRACT FROM AUTHOR]

The objective of this paper was to evaluate the accuracy of two advanced blending algorithms, Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) and Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model (ESTARFM) to downscale Moderate Resolution Imaging Spectroradiometer (MODIS) indices to the spatial resolution of Landsat. We tested two approaches: (i) "Index-then-Blend" (IB); and (ii) "Blend-then-Index" (BI) when simulating nine indices, which are widely used for vegetation studies, environmental moisture assessment and standing water identification. Landsat-like indices, generated using both IB and BI, were simulated on 45 dates in total from three sites. The outputs were then compared with indices calculated from observed Landsat data and pixel-to-pixel accuracy of each simulation was assessed by calculating the: (i) bias; (ii) R2; and (iii) Root Mean Square Deviation (RMSD). The IB approach produced higher accuracies than the BI approach for both blending algorithms for all nine indices at all three sites. We also found that the relative performance of the STARFM and ESTARFM algorithms depended on the spatial and temporal variances of the Landsat-MODIS input indices. Our study suggests that the IB approach should be implemented for blending of environmental indices, as it was: (i) less computationally expensive due to blending single indices rather than multiple bands; (ii) more accurate due to less error propagation; and (iii) less sensitive to the choice of algorithm. [ABSTRACT FROM AUTHOR]