Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “crop classifications”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 235 records · Page 13

Large Area Crop Inventory Experiment (LACIE). Accuracy assessment report phase 1A, November - December 1974

The author has identified the following significant results. Results of the accuracy assessment activity for Phase IA of LACIE indicated that (1) The 90/90 criteria could be reached if the degree of accuracy of the LACIE performance in Kansas could be equaled in other areas. (2) The classification of both wheat and nonwheat fields was significantly accurate for the three ITS segments analyzed. The wheat field classification accuracy varied for the segments. However, this was not so with respect to nonwheat fields. (3) Biophase as well as its interaction with segment location turned out to be an important factor for the classification performance. Analyst interpretation of segments for training the classifier was a significant error-contributing factor in the estimation of wheat acreage at both the field and the segment levels.

Source record↗

View angle effect in LANDSAT imagery

The view angle effect in LANDSAT 2 imagery was investigated. The LANDSAT multispectral scanner scans over a range of view angles of -5.78 to 5.78 degrees. The view angle effect, which is caused by differing view angles, could be studied by comparing data collected at different view angles over a fixed location at a fixed time. Since such LANDSAT data is not available, consecutive day acquisition data were used as a substitute: they were collected over the same geographical location, acquired 24 hours apart, with a view angle change of 7 to 8 degrees at a latitude of 35 to 45 degrees. It is shown that there is approximately a 5% reduction in the average sensor response on the second-day acquisitions as compared with the first-day acquisitions, and that the view angle effect differs field to field and crop to crop. On false infrared color pictures the view angle effect causes changes primarily in brightness and to a lesser degree in color (hue and saturation). An implication is that caution must be taken when images with different view angles are combined for classification and a signature extension technique needs to take the view angle effect into account.

Kaneko, T.↗

An assessment of LANDSAT data acquisition history on identification and area estimation of corn and soybeans

Multitemporally registered LANDSAT MSS data from four acquisitions during the 1978 growing season were used in classification of eight sample segments in Iowa and Indiana. The results illustrate that use of LANDSAT acquisition when corn has tasseled is critical, as this is the optimum time for separation of corn and soybeans. An early season acquisition when the summer crops appear as bare soil can be beneficial in reducing the confusion between these two crops and other cover types. A subset of one visible and one infrared band from each date was found to produce results not significantly different from the use of all bands. Selection of a subset of these bands may also be feasible for multitemporal analysis.

Hixson, M. M.↗

Users manual for the US baseline corn and soybean segment classification procedure

A user's manual for the classification component of the FY-81 U.S. Corn and Soybean Pilot Experiment in the Foreign Commodity Production Forecasting Project of AgRISTARS is presented. This experiment is one of several major experiments in AgRISTARS designed to measure and advance the remote sensing technologies for cropland inventory. The classification procedure discussed is designed to produce segment proportion estimates for corn and soybeans in the U.S. Corn Belt (Iowa, Indiana, and Illinois) using LANDSAT data. The estimates are produced by an integrated Analyst/Machine procedure. The Analyst selects acquisitions, participates in stratification, and assigns crop labels to selected samples. In concert with the Analyst, the machine digitally preprocesses LANDSAT data to remove external effects, stratifies the data into field like units and into spectrally similar groups, statistically samples the data for Analyst labeling, and combines the labeled samples into a final estimate.

Horvath, R.↗

Agricultural land cover mapping with the aid of digital soil survey data

A study is recounted which assessed the effect of stratifying multidate Landsat MSS data on land cover classification accuracy. The study area covered 49,184 ha (121,534 acres) in Gentry County in northwestern Missouri. A pixel-by-pixel comparison of the two land cover classifications with field-verified land cover indicated improvements in identification of all cover types when land areas were stratified by soils. The introduction of soil map information to the land cover mapping process can improve discrimination of land cover types and reduce confusion among crop types that may be caused by soil-specific management practices, soil-induced crop development differences, and background reflectance characteristics.

Stoner, E. R.↗

A simulation study of the effects of land cover and crop type on sensing soil moisture with an orbital C-band radar

Simulated C-band radar imagery for a 124-km by 108-km test site in eastern Kansas is used to classify soil moisture. Simulated radar resolutions are 100 m by 100 m, 1 km by 1 km, and 3 km by 3 km, and each is processed using more than 23 independent samples. Moisture classification errors are examined as a function of land-cover distribution, field-size distribution, and local topographic relief for the full test site and also for subregions of cropland, urban areas, woodland, and pasture/rangeland. Results show that a radar resolution of 100 m by 100 m yields the most robust classification accuracies.

Dobson, M. C.↗

Two-Stage Wildlife Event Classification for Edge Deployment

Camera-based wildlife monitoring is often overwhelmed by non-target triggers and slowed by manual review or cloud-dependent inference, which can prevent timely intervention for high stakes human–wildlife conflicts. Our key contribution is a deployable, fully offline edge vision sensor that achieves near-real-time, highly accurate wildlife event classification by combining detector-based empty-image suppression with a lightweight classifier trained with a staged transfer-learning curriculum. Specifically, Stage 1 uses a pretrained You Only Look Once (YOLO)-family detector for permissive animal localization and empty-trigger suppression, and Stage 2 uses a lightweight EfficientNet-based binary classifier to confirm puma on detector crops and gate downstream actions. Our design is robust to low-quality nighttime monochrome imagery (motion blur, low contrast, illumination artifacts, and partial-body captures) and operates using commercially available components in connectivity-limited settings. In field deployments running since May 2025, end-to-end latency from camera trigger to action command is approximately 4 s. Ablation studies using a dataset of labeled wildlife images (pumas, not pumas) show that the two-stage approach substantially reduces false alarms in identifying pumas relative to a full-image classifier while maintaining high recall. On the held-out test set (N = 1434 events), the proposed two-stage cascade achieves precision 0.983, recall 0.975, F1 0.979, accuracy 0.986, and balanced accuracy 0.983, with only 8 false positives and 12 false negatives. The system can be easily adapted for other species, as demonstrated by rapid retraining of the second stage to classify ringtails. Downstream responses (e.g., notifications and optional audio/light outputs) provide flexible actuation capabilities that can be configured to support intervention.

58 GEOSCIENCES↗

Estimation of the probability of error without ground truth and known a priori probabilities

The probability of error or, alternatively, the probability of correct classification (PCC) is an important criterion in analyzing the performance of a classifier. Labeled samples (those with ground truth) are usually employed to evaluate the performance of a classifier. Occasionally, the numbers of labeled samples are inadequate, or no labeled samples are available to evaluate a classifier's performance; for example, when crop signatures from one area from which ground truth is available are used to classify another area from which no ground truth is available. This paper reports the results of an experiment to estimate the probability of error using unlabeled test samples (i.e., without the aid of ground truth).

Havens, K. A.↗

The influence of false color infrared display on training field identification

The overall success of large-scale crop inventories of agricultural regions using Landsat multispectral scanner data is highly dependent upon the labeling of training data by analyst/photointerpreters. The principal analyst tool in labeling training data is a false color infrared composite of Landsat bands 4, 5, and 7. In this paper, this color display is investigated and its influence upon classification errors is partially determined.

Coberly, W. A.↗

User oriented ERTS-1 images

Photographic reproduction of ERTS-1 images are capable of displaying only a portion of the total information available from the multispectral scanner. Methods are being developed to generate ERTS-1 images oriented towards special users such as agriculturists, foresters, and hydrologists by applying image enhancement techniques and interactive statistical classification schemes. Spatial boundaries and linear features can be emphasized and delineated using simple filters. Linear and nonlinear transformations can be applied to the spectral data to emphasize certain ground information. An automatic classification scheme was developed to identify particular ground cover classes such as fallow, grain, rape seed or various vegetation covers. The scheme applies the maximum likelihood decision rule to the spectral information and classifies the ERTS-1 image on a pixel by pixel basis. Preliminary results indicate that the classifier has limited success in distinguishing crops, but is well adapted for identifying different types of vegetation.

Shlien, S.↗

Optimal Landsat transforms for forest applications

Eleven transformations of data from four Landsat MSS channels were investigated to find if any of the transforms accentuated the separability of natural vegetation classes in regions of varying topographical relief. Attention was given to the divergence analysis and classification accuracy of information content of the eleven transforms and four channels. A useful scaling function was observed with the second eigenvector being the denominator in the divergence values obtained. The second eigenvector was found to reduce the effects of shadowing and differential illumination of vegetation signatures, thereby enhancing the divergence values. The highest accuracies in crop identification were provided by the averages of channels 4, 6, and 7 divided by the second eigenvector.

Logan, T. L.↗

Elevation of a cane-growing area of the state of Sao Paulo using LANDSAT data

Images at a scale of 1:250.000 were visually interpreted for identification and area estimates of sugar cane plantations in Sao Paulo. The basic criteria for crop identification were the spectral characteristics of channels 5 and 7 and their temporal variations observed from different LANDSAT passes. Using this technique, it was possible to map the sugar cane areas as well as the sugar cane already harvested. An area of 801,950 hectares was estimated within the study area. The confidence interval of correct classification ranged from 87.11% to 94.71%.

Dejesusparada, N.↗

Evaluation of SLAR and thematic mapper MSS data for forest cover mapping using computer-aided analysis techniques

Supervised and cluster block training statistics were used to analyze the thematic mapper simulation MSS data (both 1979 and 1980 data sets). Cover information classes identified on SAR imagery include: hardwood, pine, mixed pine hardwood, clearcut, pasture, crops, emergent crops, bare soil, urban, and water. Preliminary analysis of the HH and HV polarized SAR data indicate a high variance associated with each information class except for water and bare soil. The large variance for most spectral classes suggests that while the means might be statistically separable, an overlap may exist between the classes which could introduce a significant classification error. The quantitative values of many cover types are much larger on the HV polarization than on the HH, thereby indicating the relative nature of the digitized data values. The mean values of the spectral classes in the areas with larger look angles are greater than the means of the same cover type in other areas having steeper look angles. Difficulty in accurately overlaying the dual polarization of the SAR data was resolved.

Hoffer, R. M.↗

Numerical trials of HISSE

The mathematical description and implementation of the statistical estimation procedure known as the Houston integrated spatial/spectral estimator (HISSE) is discussed. HISSE is based on a normal mixture model and is designed to take advantage of spectral and spatial information of LANDSAT data pixels, utilizing the initial classification and clustering information provided by the AMOEBA algorithm. The HISSE calculates parametric estimates of class proportions which reduce the error inherent in estimates derived from typical classify and count procedures common to nonparametric clustering algorithms. It also singles out spatial groupings of pixels which are most suitable for labeling classes. These calculations are designed to aid the analyst/interpreter in labeling patches with a crop class label. Finally, HISSE's initial performance on an actual LANDSAT agricultural ground truth data set is reported.

Peters, C.↗

Developing Land Use Land Cover Maps for the Lower Mekong Basin to Aid Hydrologic Modeling and Basin Planning

This paper discusses research methodology to develop Land Use Land Cover (LULC) mapsfor the Lower Mekong Basin (LMB) for basin planning, using both MODIS and Landsat satellitedata. The 2010 MODIS MOD09 and MYD09 8-day reflectance data was processed into monthlyNDVI maps with the Time Series Product Tool software package and then used to classify regionallycommon forest and agricultural LULC types. Dry season circa 2010 Landsat top of atmosphere reflectance mosaics were classified to map locally common LULC types. Unsupervised ISODATAclustering was used to derive most LULC classifications. MODIS and Landsat classifications werecombined with GIS methods to derive final 250-m LULC maps for Sub-basins (SBs) 1–8 of the LMB.The SB 7 LULC map with 14 classes was assessed for accuracy. This assessment compared randomlocations for sampled types on the SB 7 LULC map to geospatial reference data such as Landsat RGBs,MODIS NDVI phenologic profiles, high resolution satellite data, and Mekong River Commissiondata (e.g., crop calendars). The SB 7 LULC map showed an overall agreement to reference data of~81%. By grouping three deciduous forest classes into one, the overall agreement improved to ~87%.The project enabled updated regional LULC maps that included more detailed agriculture LULCtypes. LULC maps were supplied to project partners to improve use of Soil andWater AssessmentTool for modeling hydrology and water use, plus enhance LMB water and disaster managementin a region vulnerable to flooding, droughts, and anthropogenic change as part of basin planningand assessment.

land use land cover mapping; SWAT hydrologic model↗

Land management effects on wet aggregate stability and carbon content

Land management affects soil structure and many other soil properties and processes. Our objectives were to evaluate soil organic C (SOC), aggregate size distribution, aggregate-associated C, and soil structure as affected by long-term land management and slope. A chronosequence of 38 on-farm sites with low to high (5–18%) slopes was selected to evaluate 5–40 yr of management. The sites were classified as business as usual (BAU) cropland (BAU-Crop), BAU pasture (BAU-Past), newly established conservation reserve program (CRP) areas (CRP-New), and established CRP (CRP-Old). Soil samples were collected from the 0-to-5- and 5-to-15-cm depth increments and processed for soil property measurements including fractionation by wet sieving into five aggregate size classes (>2,000, 1,000–2,000, 500–1,000, 250–500, and 53–250 μm). Within the surface 5 cm, mean weight diameter (MWD) and geometric mean diameter (GMD) were used to characterize soil structural stability. The BAU-Past and CRP-Old sites had 79% more macroaggregates (>2,000, 1,000–2,000, and 500–1,000 μm), 123% higher MWD, 38% higher GMD, and 47% higher SOC than BAU-Crop or CRP-New sites. The 5-to-15-cm depth increment showed a similar but lower magnitude response. Aggregate-associated C was quantified using a constant soil mass that reflected aggregate size distribution to prevent overestimating C content. Lower-slope locations had more SOC, more macroaggregates, more C associated with macroaggregates, and higher GMD and MWD compared with high-slope locations across all management classifications and soil depths. The results support our hypothesis that the high-slop soils may benefits from specific management decisions than the lower-sloping soils as a function of landscape property. We recommend reestablishing grassland on sloping land that is susceptible to excessive soil erosion, although those practices will likely take a long time to restore soil structural stability and SOC content to precultivation levels.

54 ENVIRONMENTAL SCIENCES↗

A study of Quaternary landforms and materials in the Midwest and Great Plains

The author has identified the following significant results. Measurements made from prints of ERTS-1, MSS 5 images, show practical limits of detectability for this imagery in the Midwest. The smallest high contrast object detectable has an approximate measured diameter of 150 feet. The smallest clearly identifiable cultural feature is roughly 300 feet for high contrast, and 400 to 500 feet for low contrast objects. Rural roadways, with an average width of 75 feet, are clearly defined due to high reflectivity, linearity, and the instantaneous field of view of the scanner. On the infrared a farm pond slightly greater than one acre is detectable. Crop and natural foliage cover in the Midwest during summar months obscures geologic and soils information and hinders detailed mapping. In the western Great Plains large-scale mapping of this kind may be possible, even at this time of year. In southwestern Iowa, topographic and drainage system anomalies, revealed by the imagery, are related to the slope of and depth to the buried bedrock surface. In eastern Iowa land use classification can be done from ERTS-1 imagery.

Morrison, R. B.↗

Quantifying leaf symptoms of sorghum charcoal rot in images of field‐grown plants using deep neural networks

Abstract Charcoal rot of sorghum (CRS) is a significant disease affecting sorghum crops, with limited genetic resistance available. The causative agent, Macrophomina phaseolina (Tassi) Goid, is a highly destructive fungal pathogen that targets over 500 plant species globally, including essential staple crops. Utilizing field image data for precise detection and quantification of CRS could greatly assist in the prompt identification and management of affected fields and thereby reduce yield losses. The objective of this work was to implement various machine learning algorithms to evaluate their ability to accurately detect and quantify CRS in red‐green‐blue images of sorghum plants exhibiting symptoms of infection. EfficientNet‐B3 and a fully convolutional network emerged as the top‐performing models for image classification and segmentation tasks, respectively. Among the classification models evaluated, EfficientNet‐B3 demonstrated superior performance, achieving an accuracy of 86.97%, a recall rate of 0.71, and an F1 score of 0.73. Of the segmentation models tested, FCN proved to be the most effective, exhibiting a validation accuracy of 97.76%, a recall rate of 0.68, and an F1 score of 0.66. As the size of the image patches increased, both models’ validation scores increased linearly, and their inference time decreased exponentially. This trend could be attributed to larger patches containing more information, improving model performance, and fewer patches reducing the computational load, thus decreasing inference time. The models, in addition to being immediately useful for breeders and growers of sorghum, advance the domain of automated plant phenotyping and may serve as a foundation for drone‐based or other automated field phenotyping efforts. Additionally, the models presented herein can be accessed through a web‐based application where users can easily analyze their own images.

Gonzalez, Emmanuel M.↗