Search

When sufficient time-series images and training data are unavailable for crop classification, features extracted from convolutional neural network (CNN)-based representative learning may not provide useful information to discriminate crops with similar spectral characteristics, leading to poor classification accuracy. In particular, limited input data are the main obstacles to obtain reliable classification results for early crop mapping. This study investigates the potential of a hybrid classification approach, i.e., CNN-random forest (CNN-RF), in the context of early crop mapping, that combines the automatic feature extraction capability of CNN with the superior discrimination capability of an RF classifier. Two experiments on incremental crop classification with unmanned aerial vehicle images were conducted to compare the performance of CNN-RF with that of CNN and RF with respect to the length of the time-series and training data sizes. When sufficient time-series images and training data were used for the classification, the accuracy of CNN-RF was slightly higher or comparable with that of CNN. In contrast, when fewer images and the smallest training data were used at the early crop growth stage, CNN-RF was substantially beneficial and the overall accuracy increased by maximum 6.7%p and 4.6%p in the two study areas, respectively, compared to CNN. This is attributed to its ability to discriminate crops from features with insufficient information using a more sophisticated classifier. The experimental results demonstrate that CNN-RF is an effective classifier for early crop mapping when only limited input images and training samples are available.

In 1980s, Crop Water Stress Index (CWSI) is suggested to indicate the water stress of crops. CWSI is based on the leaf energy balance, which is closely related to leaf temperature. To calculate CWSI, meteorological factors such as air temperature and vapor pressure deficit should be measured besides leaf temperature. As recent technology has been developed, leaf temperature can be easily observed by thermal camera or infrared thermometer. Stomatal conductance (gs, mmol m-2 s-1) is one of the critical factors to understand crop photosynthesis and water demand. In addition, the behaviors of gs can represent the biotic and abiotic plant stresses. In abnormal condition, such as drought, insects or disease, gs getting lower. The observation of gs will make better to evaluate and predict crop growth and conditions. Therefore, the time series data of gs is useful for the monitoring of crop growth and the quick detection of abnormal crop condition in smart-farming system but there are some limitations to measure gs continuously and easily. We assume that there is some relationship between CWSI and gs because both has strong relation to leaf temperature. Thus, the aim of this study is to investigate possibility of estimation of gs using CWSI which is derived from thermal image. Through the data collected from literatures, negative correlations between CWSI and gs were revealed. The slope of correlation was changed according to crop types. In addition, as a result of simulation, there is almost linear negative relationship between CWSI and gs, and the slope was determined by maximum stomatal conductance (gs_max). Field measurement in this study was also demonstrated to identify such correlation. Further, various methods to measure CWSI were tested. This relationship will contribute to not only monitoring of crop stress for irrigation scheduling in smart farm system but also estimating evapotranspiration, photosynthesis, and crop yield.

Recently, usage of UAV (Unmanned Aerial Vehicle) has increased in agricultural part. This study was conducted to classify onion and garlic using supervised classification of a fixed-wing UAV (Model : Ebee) images for evaluation of possibility about estimation of onion and garlic cultivation area using UAV images. Aerial images were obtained 11~12 times from study sites in Changryeng-gun and Hapcheon-gun during farming season from 2015 to 2016. The result for accuracy in onion and garlic image classification by R-G-B and R-G-NIR images showed highest Kappa coefficients for the maximum likelihood method. The result for accuracy in onion and garlic classification showed high Kappa coefficients of 0.75~0.97 from DOY 105 to DOY 141, implying that UAV images could be used to estimate onion and garlic cultivation area.

Recent technological advance in UAV (Unmanned Aerial Vehicle) technology offers new opportunities for assessing crop situation using UAV imagery. The objective of this study was to assess if reflectance and NDVI derived from consumer-grade cameras mounted on UAVs are useful for crop condition monitoring. This study was conducted using a fixed-wing UAV(Ebee) with Cannon S110 camera from March 2015 to March 2016 in the experiment field of National Institute of Agricultural Sciences. Results were compared with ground-based recordings obtained from consumer-grade cameras and ground multi-spectral sensors. The relationship between raw digital numbers (DNs) of UAV images and measured calibration tarp reflectance was quadratic. Surface (lawn grass, stairs, and soybean cultivation area) reflectance obtained from UAV images was not similar to reflectance measured by ground-based sensors. But NDVI based on UAV imagery was similar to NDVI calculated by ground-based sensors.

For more than 50 years, satellite images have been used to monitor crop growth. Currently, unmanned aerial vehicle (UAV) imagery is being assessed for analyzing within field spatial variability for agricultural precision management, because UAV imagery may be acquired quickly during critical periods of rapid crop growth. This study refers to the derivation of growth estimating equation for highland Kimchi cabbage using UAV derived normalized difference vegetation index (NDVI) and agro-meteorological factors. Anbandeok area in Gangneung, Gangwon-do, Korea is one of main districts producing highland Kimchi cabbage. UAV imagery was taken in the Anbandeok ten times from early June to early September. Meanwhile, three plant growth parameters, plant height (P.H.), leaf length (L.L.) and outer leaf number (L.N.), were measured for about 40 plants (ten plants per plot) for each ground survey. Six agro-meteorological factors include average temperature; maximum temperature; minimum temperature; accumulated temperature; rainfall and irradiation during growth period. The multiple linear regression models were suggested by using stepwise regression in the extraction of independent variables. As a result, NDVIUAV and rainfall in the model explain 93% of the P.H. and L.L. with a root mean square error (RMSE) of 2.22, 1.90 cm. And NDVIUAV and accumulated temperature in the model explain 86% of the L.N. with a RMSE of 4.29. These lead to the result that the characteristics of variations in highland Kimchi cabbage growth according to NDVIUAV and other agro-meteorological factors were well reflected in the model.

Italian Ryegrass (IRG), which is known as high yielding and the highest quality winter annual forage crop, is grown in mid-south area in Korea. The objective of this study was to evaluate the use of unmanned aerial vehicle (UAV) for the monitoring IRG growth. Unmanned aerial vehicle imagery obtained from middle March to late May in Nonsan, Chungcheongnam-do. Unmanned aerial vehicle imagery corrected geometrically and atmospherically to calculate normalized difference vegetation index (NDVI). We analyzed the relationships between NDVIUAV of IRG and biophysical measurements such as plant height, fresh weight, and dry weight over an entire IRG growth period. The similar trend between NDVIUAV and growth parameters was shown. Correlation analysis between NDVIUAV and IRG growth parameters revealed that NDVIUAV was highly correlated with fresh weight (r=0.988), plant height (r=0.925), and dry weight (r=0.853). According to the relationship among growth parameters and NDVIUAV, the temporal variation of NDVIUAV was significant to interpret IRG growth. Four different regression models, such as (1) Linear regression function, (2) Linear regression through the origin, (3) Power function, and (4) Logistic function were developed to evaluate the relationship between temporal NDVIUAV and measured IRG growth parameters. The power function provided higher accurate results to predict growth parameters than linear or logistic functions using coefficient of determination. The spatial distribution map of IRG growth was in strong agreement with the field measurements in terms of geographical variation and relative numerical values when NDVIUAV was applied to power function. From these results, NDVIUAV can be used as a new tool for monitoring IRG growth.

Kimchi cabbage is one of the most important vegetables in Korea and a target crop for market stabilization as well. In particular Kimchi cabbages in a highland area are very sensitive to the fluctuations in supply and demand. Yield variability due to growth conditions dictates the market fluctuations of Kimchi cabbage price. This study was carried out to understand the distribution of the highland Kimchi cabbage growth status in Anbandeok. Anbandeok area in Gangneung, Gangwon-do, Korea is one of the main producing districts of highland Kimchi cabbage. The highland Kimchi cabbage status map of each growth factor was obtained from unmanned aerial vehicle (UAV) imagery and field survey data. Six status maps include UAVRGB image map, normalized difference vegetation index (NDVI) distribution/anomaly map, Crop distribution map, Planting/Harvest distribution map, Growth parameter map and Growth disorder map. As a result, the highland Kimchi cabbage status maps from May 31 to Sep. 6 in 2016 were presented to show spatial variability in the field. The benefits of the highland Kimchi cabbage status map can be summarized as follows: crop growth monitoring, reference for field observations and survey, the relative comparison of the growth condition in field scale, evaluation of growth in comparison of average year, change detection of annual crops or planting areas, abandoned fields monitoring, prediction of harvest season etc.

Unmanned Aerial Vehicle (UAV) imagery are being assessed for analyzing within field spatial variability for agricultural precision management, because unmanned aerial vehicle imagery may be acquired quickly during critical periods of rapid crop growth. The objective of this study was to evaluate the use of unmanned aerial vehicle for the monitoring barley growth. Unmanned aerial vehicle imagery obtained from middle February to late June in Wanju, Jeollabuk-do. Unmanned aerial vehicle imagery corrected geometrically and atmospherically to calculate normalized difference vegetation index (NDVI). We analyzed the relationships between NDVI UAV of barley and biophysical measurements such as plant height, number of tiller and shoot dry weight over an entire barley growth period. The similar trend between NDVI UAV and growth parameters was shown. Correlation analysis between NDVI UAV and barley growth parameters revealed that NDVI UAV was highly correlated with shoot dry weight (r=0.932) and plant height (r=0.879). According to the relationship among growth parameters and NDVI UAV , the temporal variation of NDVIUAV was significant to interpret barley growth. The spatial distribution map of barley growth was in strong agreement with the field measurements in terms of geographical variation and relative numerical values when NDVI UAV was applied to power function. From these results, NDVI UAV can be used as a new tool for monitoring barley growth.

For more than 50 years, satellite images have been used to monitor crop growth. Currently, unmanned aerial vehicle (UAV) imagery are being assessed for analyzing within field spatial variability for agricultural precision management, because UAV imagery may be acquired quickly during critical periods of rapid crop growth. This study refers to the derivation of highland Kimchi cabbage growth prediction equation by using UAV derived normalized difference vegetation index (NDVI) and agro-meteorological factors. Anbandeok area in Gangneung, Gangwon-do, Korea is one of main producing district of highland Kimchi cabbage. UAV imagery was taken on the Anbandeok twelve times from early June to early September during the highland Kimchi cabbage growing season. Meanwhile, field reflectance spectra and three plant growth parameters, including plant height (P.H.), leaf height (L.H.) and leaf number (L.N.), were measured for about 40 plants (ten plants per plot) for each ground survey. The NDVIField for each of the 40 plants was measured using an active plant growth sensor (Crop CircleTM) at the same time. Six agro-meteorological factors include average temperature; maximum temperature; minimum temperature; accumulated temperature; rainfall and irradiation. The multiple linear regression models were suggested by using stepwise regression in the extraction of independent variables. Of the output independent variables of the stepwise regression, NDVIUAV performs best (r=0.933). And correlation coefficients of highland Kimchi cabbage growth parameters, accumulated temperature, irradiation and rainfall showed positive relationship; the results were respectively 0.931, 0.926, 0.893 (p

Soil moisture content is generally accepted as an important factor to understand the process of crop growth and is the basis of earth system models for analysis and prediction of the crop condition. To continuously monitor soil moisture changes at kilometer scale, it is demanded to create high resolution data from the current, several tens of kilometers. In this paper we described a downscaling method for Advanced Microwave Scanning Radiometer 2 (AMSR2) Soil Moisture Content (SMC) from 10 km to 30 m resolution using a soil texture and field measurements that have a high correlation with the SMC. As a result, the soil moisture variations of both data (before and after downscaling) were identical, and the Root Mean Square Error (RMSE) of SMC exhibited the low values. Also, time series analyses showed that three kinds of SMC data (field measurement, original AMSR2, and downscaled AMSR2) had very similar temporal variations. Our method can be applied to downscaling of other soil variables and can contribute to monitoring small-scale changes of soil moisture by roviding high resolution data.

Temporal and spatial variation of soil moisture is important for understanding patterns of climate change, for developing and evaluating land surface models, for designing surface soil moisture observation networks, and for determining the appropriate resolution for satellite-based remote sensing instruments for soil moisture. In this study, we measured several soil moistures in upland soil using Advanced Microwave Scanning Radiometer 2 (AMSR2) Soil Moisture Content (SMC) during eight-month period in Chungbuk province. The upland soil moisture properties were expressed by simple statistical methods (average, standard deviation and coefficient of variation) from the monthly context. Supplementary studies were also performed about the effect of top soil texture on the soil moisture responses. If the results from this study were utilized well in specific cities and counties in Korea, it would be helpful to establish the countermeasures and action plans for preventing disasters because it was possible to compare with the relationship between soil moisture and top soil texture of each region. And it would be the fundamental data for estimating the effect of future agricultural plan.

The NDVI (Normalized difference vegetation index) is used as indicators of crop growth situation in remote sensing. To measure or validate the NDVI, reliable NDVI sensors have been needed. We tested new fixed-field NDVI sensor, “SRS (Spectral Reflectance Sensor)” developed by Decagon Devices, during Kimchi cabbage growing season at the cultivation area located in Gochang, Gangneung and Taebaek in Korea from 2014 to 2015. The diurnal variation of NDVI measured by SRS (SRS NDVI) showed a slight ∩-profile shape and was affected by water on the sensor surface. This means that SRS NDVI around noontime is resonable, except rainy day. Comparisons were made between the SRS NDVI and NDVI of used widely mobile sensor (Cropcircle NDVI). The comparisons indicate that SRS NDVI are close to Cropcircle NDVI (R=0.99). SRS NDVI time series displayed change of the plant height and leaf width of Kimchi cabbage. An obvious exponential relationship is found between SRS NDVI and the plant height (R 2 ≥0.92) and leaf width(R 2 ≥0.92) of Kimchi cabbage. Thus, SRS NDVI will be used as indicator of crop growth situation and a very powerful tool for evaluation of remote sensing NDVI estimates and associated corrections.

We present a novel GA based feature extractor for network optimal initialization.Our method selects more dominant feature extractor in merge phase using GA.Results show improvements recognition performance in comparison with DBNs.We also suggest a new approach for retraining additional classes as its application.Our approach for retraining can add output classes at lower error rate than DBNs. In this paper, we propose a novel split training and merge algorithm for deep learning. The proposed algorithm improves recognition accuracy and suggests a new approach for retraining. The algorithm is motivated by the genetic algorithm (GA) and is composed of two procedures. The first procedure initializes two individual networks using deep belief networks (DBNs), and the second procedure merges the two networks using the GA. Biases and weights of the network that is trained using DBNs are represented as a matrix between each layer, and each row of this matrix is used as a chromosome in the merge procedure. To evaluate the performance, we conduct two set of experiments. The first set is to recognize accuracy of the proposed algorithm, and the second set is for a new retraining approach. The results show that the proposed algorithm has a lower average error rate (6.84 ? 4.57%) than the DBNs, and it can add classes at a lower average error rate (9.06 ? 6.17% and 10.17 ? 4.51%) without pre-training using the restrict Boltzmann machines (RBMs) for existing classes data.

Chinese cabbage is one of the most important vegetables in Korea and a target crop for market stabilization as well. Remote sensing has long been used as a tool to extract plant growth, cultivated area and yield information for many crops, but little research has been conducted on Chinese cabbage. This study refers to the derivation of simple Chinese cabbage growth prediction equation by using RapidEye derived vegetation index. Daesan-myeon area in Gochang-gun, Jeollabuk-do, Korea is one of main producing district of Chinese cabbage. RapidEye multi–spectral imagery was taken on the Daesan-myeon five times from early September to late October during the Chinese cabbage growing season. Meanwhile, field reflectance spectra and five plant growth parameters, including plant height (P.H.), plant diameter (P.D.), leaf height (L.H.), leaf length (L.L.) and leaf number (L.N.), were measured for about 20 plants (ten plants per plot) for each ground survey. The normalized difference vegetation index (NDVI) for each of the 20 plants was measured using an active plant growth sensor (Crop Circle TM ) at the same time. The results of correlation analysis between the vegetation indices and Chinese cabbage growth data showed that NDVI was the most suited for monitoring the L.H. (r=0.958~0.978), L.L. (r=0.950~0.971), P.H. (r=0.887~0.982), P.D. (r=0.855~0.932) and L.N. (r=0.718~0.968). Retrieval equations were developed for estimating Chinese cabbage growth parameters using NDVI. These results obtained using the NDVI is effective provided a basis for establishing retrieval algorithm for the biophysical properties of Chinese cabbage. These results will also be useful in determining the RapidEye multi-spectral imagery necessary to estimate parameters of Chinese cabbage.

To understand the impact of 2015 spring drought on crop production of DPRK (Democratic People’s Republic of Korea), we analyzed satellite and weather data to produce 2015 spring outlook of rice paddy field and rice growth in relation to weather anomaly. We defined anomaly of 2015 for weather and NDVI in comparison to past 5 year-average data. Weather anomaly layers for rainfall and mean temperature were calculated based on 27 weather station data. Rainfall in late April, early May, and late May in 2015 was much lower than those in average years. NDVI values as an indicator of rice growth in early June of 2015 was much lower than in 2014 and the average years. RapidEye and Radarsat-2 images were used to monitor status of rice paddy irrigation and transplanting. Due to rainfall shortage from late April to May, rice paddy irrigation was not favorable and rice planting was not progressed in large portion of paddy fields until early June near Pyongyang. Satellite images taken in late June showed rice paddy fields which were not irrigated until early June were flooded, assuming that rice was transplanted after rainfall in June. Weather and NDVI anomaly data in regular basis and timely acquired satellite data can be useful for grasping the crop and land status of DPRK, which is in high demand.

Remote sensing can be used to provide information about the monitoring of crop situation. This study was conducted to estimate the amount of nitrogen present in paddy fields by measuring the amount of nitrogen in hairy vetch using an UAV (Unmaned Aerial Vehicle). NDVIs (Normalized Difference Vegetation Index) were calculated using UAV images obtained from paddy fields in Seocheon on May 14 th 2015. There was strong relationship between UAV NDVI and the amount of nitrogen in hairy vetch (R 2 =0.79). Spatial distribution maps of green manure nitrogen were generated on each paddy field using the nitrogen-vegetation index relations to help farmers determine the amount of N fertilizers added to their rice fields after the application of green manure such as hairy vetch.

Unmanned Aerial Vehicle (UAV) imagery are being assessed for analyzing within field spatial variability for agricultural precision management, because unmanned aerial vehicle imagery may be acquired quickly during critical periods of rapid crop growth. The objective of this study was to evaluate the use of unmanned aerial vehicle for the monitoring barley growth. Unmanned aerial vehicle imagery obtained from middle February to late June in Wanju, Jeollabuk-do. Unmanned aerial vehicle imagery corrected geometrically and atmospherically to calculate normalized difference vegetation index (NDVI). We analyzed the relationships between NDVI UAV of barley and biophysical measurements such as plant height, number of tiller and shoot dry weight over an entire barley growth period. The similar trend between NDVIUAV and growth parameters was shown. Correlation analysis between NDVIUAV and barley growth parameters revealed that NDVIUAV was highly correlated with shoot dry weight (r=0.932) and plant height (r=0.879). According to the relationship among growth parameters and NDVIUAV, the temporal variation of NDVIUAV was significant to interpret barley growth. The spatial distribution map of barley growth was in strong agreement with the field measurements in terms of geographical variation and relative numerical values when NDVIUAV was applied to power function. From these results, NDVIUAV can be used as a new tool for monitoring barley growth.

The remote monitoring system of water salinity was assessed in Wando reclaimed land lake during a farming season in 2009. Increasing of water salinity in this lake used to bring about salt damage on rice plant occasionally. At the early stage of the rice growing period, rice growth was not damaged due to enough rainfall with more than 120 mm from the mid-May to the first ten days of June. Data collection using on-site water salinity measuring sensors every 2 hours and real-time transmission in system were carried out for the experiment. We compared the transmitted values from the sensor system with water sample values collected and analyzed by a local technical office. Salt concentrations measured by sensor in real-time monitoring system were available data. The regression equation between rainfall and water salinity was presented as(water salinity after rainfall) = 0.621×(water salinity before rainfall)×exp(-0.0139×rainfall), (r 2 =0.579, p

Radar remote sensing is appropriate for rice monitoring because the areas where this crop is cultivated are often cloudy and rainy. Especially, Synthetic Aperture Radar (SAR) can acquire remote sensing information with a high temporal resolution in tropical and subtropical regions due to its all-weather capability. This paper analyzes the relationships between backscattering coefficients of rice measured by RADARSAT-2 SAR and growth parameters during a rice growth period. And we applied the relationships to crop monitoring of paddy rice in North Korea. As a result, plant height and Leaf Area Index (LAI) increased until Day Of Year (DOY) 234 and then decreased, while fresh weight and dry weight increased until DOY 253. Correlation coefficients revealed that Horizontal transmit and Horizontal receive polarization (HH)-polarization backscattering coefficients were correlated highly with plant height (r

A ground-based microwave scatterometer has an advantage for monitoring soil moisture content using multi-polarization, multi-frequencies and various incidence angles. In this paper, ground-based multi-frequency (L-, C-, and X-band) polarimetric scatterometer system capable of making observations every 10 min was used to monitor the soil moisture conditions in a corn field over an entire growth cycle. Measurements of volumetric soil moisture were obtained and their relationships to the backscatter observations were examined. Time series of soil moisture content was not corresponding with backscattering coefficient pattern over the whole growth stage, although it increased until early July (Day Of Year, DOY 160). We examined the relationship between the backscattering coefficients from each band and soil moisture content of the field. Backscattering coefficients for all bands were not correlated with soil moisture content when considered over the entire stage (r≤0.48). However, L-band Horizontal transmit and Horizontal receive polarization (HH) had a good correlation with soil moisture (r=0.85**) when LAI was lower than 2. Prediction equations for soil moisture were developed using the L-HH data. Relation between L-HH and soil moisture shows linear pattern and related with soil moisture content (R²=0.77). Results from this study show that backscattering coefficients of microwave scatterometer appear to be effective to estimate soil moisture content in the field level.

Ground-based polarimetric scatterometers have been effective tools to monitor the growth of crop with multi-polarization and frequencies and various incident angles. An important advantage of these systems that can be exploited is temporal observation of a specific crop target. Polarimetric backscatter data at L-, C- and X-bands were acquired every 10 minutes. We analyzed the relationships between L-, C- and X-band signatures, biophysical measurements over the whole corn growth period. The Vertical transmit and Vertical receive polarization (VV) backscattering coefficients for all bands were greater than those of the Horizontal transmit and Horizontal receive polarization (HH) until early-July, and then thereafter HH-polarization was greater than VV-polarization or Horizontal transmit and Vertical receive polarization (HV) until the harvesting stage (Day Of Year, DOY 240). The results of correlation analysis between the backscattering coefficients for all bands and corn growth data showed that L-band HH-polarization (L-HH) was the most suited for monitoring the fresh weight (r=0.95***), dry weight (r=0.95***), leaf area index (r=0.86**), and vegetation water content (r=0.93***). Retrieval equations were developed for estimating corn growth parameters using L-HH. The results indicated that L-HH could be used for estimating the vegetation biophysical parameters considered here with high accuracy. Those results can be useful in determining frequency and polarization of satellite Synthetic Aperture Radar stem and in designing a future ground-based microwave system for a long-term monitoring of corn.