Condition assessment of pearl millet/ bajra crop in different vigour zones using Radar Vegetation Index

Abstract

The study was focused on assessing the condition of pearl millet crop in critical growth stages using both polarimetric Radarsat-2 and dual-polarized Sentinel-1 datasets. The results revealed that bajra having a close structured phenology like maize and Jowar, exhibited significant changes in RVI due to differences in the crop calendar dates. For bajra, polarimetric RVI generated from information rich Radarsat-2 was observed to have a higher level of saturation till 6 kgm−2 biomass with a R2 of 0.7. In all instances, RVI exhibited a significant relationship with VWC and plant volume with a R2 above 0.7 due to its higher sensitivity towards crop dielectric constant. Unlike NDVI, RVI increased with an increase in Leaf Area Index till 5.8 even during panicle initiation stage. Backscatter and truncated RVI almost follow a similar trend of RVI response for various crop growth parameters. Hence, in case of regional analysis and high cost of Radarsat-2 dataset, one can use freely available sentinel data for RVI analysis due to its wider swath coverage. The observed high correlation of crop age with RVI (R2 = 0.6) proved to be the best tool for predicting sowing dates in staggered sowing zones.

Introduction

Microwave remote sensing has the ability to provide cloud free data during monsoon season which helps in continuous and efficient monitoring of crops in cloud persistent areas. Moreover unique sensitivity of radar waves to vegetation structure, shape, size, vegetation orientation and dielectric constant yields complementary information than optical remote sensing which takes account of only spectral reflectance of crops [1,2,3]. The dielectric content and structure of vegetation varies as crop undergoes various growth stages starting from early vegetative to senescence and capturing these temporal changes is the most important aspect for identifying different crops [4, 5]. The study of dynamic crop growth parameters through radar observations for wide range of system configurations always remains a challenge. Lot of previous studies were taken up to investigate the microwave sensor response to various biophysical parameter of crops and its potential in crop condition assessment [6,7,8], additional studies should be extended to variety of crops for developing robust retrieval methodologies. In optical remote sensing, various studies established relationship between Normalized Difference Vegetation Index (NDVI) and various crop parameters like Vegetation Water Content (VWC) and Leaf Area Index (LAI) for characterizing vegetation canopy [8, 9]. It reveals that SAR variables derived from satellite data are sensitive to crop parameters which aids a tool in crop condition assessment [10].

Kim and Van Zyl (2001) proposed the Radar Vegetation Index (RVI) (which has lesser environmental effects) for polarimetric SAR data, which was used for estimating the forest biomass and characterizing various forest parameters[11]. Later it was extended to estimate VWC of rice and soybean crop in multi-frequency bands [12]. VWC finds an important application in retrieving soil moisture from remote sensing data [13] and also useful in drought monitoring. Several authors reported the significant correlation of RVI with various crop biophysical parameters. For wheat crop, RVI shows high sensitivity to VWC and wet biomass than single channel backscatter coefficients, as vegetation structure were more complex and random [14]. RVI proved to have higher sensitivity to forest biomass, where L-band provides better results in characterizing biomass than X and C-bands. Higher level of biomass saturation up to 180–200 ton/ha was noted in L-band because of their ability to penetrate the forest canopy, where C-band saturates faster at 120–140 ton/ha [15]. Haldar et.al (2018) proposed the truncated RVI equation for dual- polarized datasets by assuming VV backscattering was equal to HH in mid to advanced crop growth stages. The condition of cotton crop was monitored by establishing various polynomial relationship of RVI with biomass, VWC and crop age having a greater dynamic range and the results were verified using RISAT-1 compact polarimetric datasets [16].

According to our knowledge, no previous work related to RVI have been reported in coarse cereals for monitoring growth conditions. The major focus of the current study is to apply the proposed RVI [11] to study the crop growth condition, different growth stages and vigour assessment of bajra crop. This was done to analyze how well crop parameters correlated with RVI in terms of dryland crop monitoring. Moreover, pearl millet (also known as bajra) being a rainfed crop, generally crop stress or condition assessment serves as a vital input in management practices like fertilizer application and irrigation. The novelty of the research is to establish relationship between polarimetric and backscatter RVI derived from C-band Radarsat-2 quad polarized data with ground measured key biophysical parameters like wet biomass, VWC, crop age, plant height, LAI, plant volume and plant density. The same relationship was attempted with truncated RVI [16] obtained from VV + VH polarizations of Sentinel-1 data having wider swath. This was done to test the applicability and potential of Sentinel-1 data for regional crop studies. The study was done in a robust way to find operational technique to characterize various crop parameters in varied vigour zones. This technique tries to minimize the impact arise due to incidence angle, crop structure and environmental conditions because of its low sensitivity to environment condition effects. We also studied the temporal RVI response of other crops in addition to bajra for monitoring crop condition and RVI changes in different growth stages. Time series RVI relationship was established from July to September to find the workable range of biomass for bajra crop in the C-band SAR dataset. RVI depends on incidence angle and path length, where previous studies used shallow incidence angle for crop studies as it increases the path length through vegetation and increase vegetation response [17, 18]. Hence, the current study was also focused on using higher incidence angle in both Radarsat-2 and Sentinel-1 dataset. This study comparatively analyzes the potential of polarimetric RVI, backscatter RVI from Radarsat-2 and truncated RVI from Sentinel-1 datasets for dryland crop monitoring and vigour assessment.

Materials and methods

Study site

The study investigates the parts of Agra district located in Uttar Pradesh covering an area of about 4027 km2, which was selected as one of the highest contributors (Fig. 1). The area is bounded by 27010′59′’ N latitude 78001′00′’E longitude, with an average elevation of 166 m above m.s.l. The monsoon starts from the month of June and continues till October. The average annual rainfall (2009–2017) was 655.5 mm. The overall trend shows that there was a decrease in rainfall over years from 2010 to 2017 and 2018 received high rainfall of above normal (728.9 mm) (Indian Meteorological Dept.)

Fig. 1
figure1

Ground truth points overlaid on three date (24th July, 17th Aug and 10th Sep) False Color Composite image generated from HH polarization of Radarsat-2 datasets

Pearl millet (bajra) is the dominant kharif crop grown in this region from June to October. Paddy, which is the second dominant crop in low lying areas was transplanted in the mid of July and early August. Jowar and maize were sown during end of May and almost reaches harvesting stage in August. In some parts, jowar is late sown around end of June. Sowing of bajra crop starts from the end of June and continues till end of July. Among cotton, Bt cotton was sown early during May and indigenous cotton was sown in the month of June. High biomass crops like sesbania and arhar comprises the minor crops.

Satellite data acquired

The satellite dataset were acquired in synchronous with various critical stages of pearl millet crop. Radarsat-2 quad polarized dataset acquired by Canadian Space agency from July to September, 2018 was used in the study. The truncated RVI obtained from Sentinel-1 data holds well with backscatter RVI of Radarsat-2 only in mid to advanced growth stages as interaction of radar waves between soil and vegetation is negligible. Hence, July Sentinel-1 dataset has been excluded while August and September datasets coinciding with mid stages were taken into account for RVI analysis. Five date Sentinel-1A GRD dataset of 12 days temporal resolution were downloaded from Copernicus Data hub provided by European Space Agency (Table 1).

Table 1 SAR data specifications used in the study

Ground truth data collection

Ground truth data were collected in synchronous to satellite pass. The coordinates of the crop field were marked with GPS (Global Positioning System) receiver. Generally, crop fields more than 3 to 5 ha over a continuous stretch were selected for sampling. Plant parameters like crop sowing/transplanted date, crop stage, crop vigor, crop height, crop percentage cover, leaf Area Index (LAI) of different crops and soil parameters like soil moisture and roughness were collected as shown in Table 2.

Table 2 Crop/soil parameters collected

Stratified random sampling of more than 180 points for various Bajra categories (high and low biomass) were sampled due to observed high variability and were surveyed three times to observe the changes in crop phenology and field water logging conditions. 30 paddy points and 20 points per other crops were randomly sampled. Ground truth campaign was conducted on early-vegetative, peak vegetative and grain filling stages of Bajra crop as shown in Fig. 2.

Fig. 2
figure2

Critical growth phases of pearl millet crop: a Early vegetative stage, b Peak vegetative stage, and c Panicle Initiation stage

Wet Biomass (above ground biomass) samples for each field was collected in different growth stages and weighed to get fresh weight of Bajra crop. Later the samples were oven dried in laboratory to get dry weight. The plant volume of Bajra crop was estimated using Archimedes principle [19]. Vegetation Water Content was estimated by differencing wet biomass (field measured) and dry Biomass (processed in laboratory) [12].

Extraction of radar vegetation index

Both Radarsat -2 and Sentinel- 1 were pre-processed using Sentinel-1 Toolbox developed by European Space Agency (ESA). The orbit file correction was carried out for Sentinel-1 dataset to update the metadata with precise orbit vectors. Radiometric calibration was done to adjust the radar backscatter values so that it represents radar cross section of the ground targets. The study area consists of mostly agricultural areas with varied crops leading to random scattering process. For such area, presently adaptive filter Lee Sigma 5 × 5 window size [20] was found to be effective for the particular scene as it removes the speckles without compromising the edge sharpness, scattering mechanisms and preserves finer image details.

Previous studies indicate the use of radar-based variable RVI for vegetation condition monitoring. Generally water bodies and urban features shows RVI near 0 and for crop RVI increases with increase in crop growth (peak vegetative stage) and dips when crop enters into reproductive stage due to yellowing of leaves and decreasing of vegetation water content [12]. For forest, previous studies shows that RVI may go beyond 1 [15].

Studies were conducted to investigate the relation and sensitivity of radar parameters with various biophysical parameters of crops [21, 22]. Small scale variation can be quantified by the use of high-resolution radar systems. It was observed that backscatter response of VV polarization was equal to HH polarization in the mid to advanced crop growth stages[16]. The cross polarization component HV was taken into account in estimating RVI as vegetation is highly sensitive to cross polarized response capturing significant volume scattering and multiple scattering information from crops [22]. In this present study, truncated RVI (as given in Eq. 1) was generated from dual polarized Sentinel-1 VV + VH data for analyzing temporal RVI response of various crops and its relationship with various biophysical parameters of bajra in advanced crop growth stages.

$$ {\text{Truncated}} \, {\text{RVI}} = \frac{{4.\sigma_{VH}^{0} }}{{\sigma_{VV}^{0} + \sigma_{VH}^{0} }} $$
(1)

The results were verified using high resolution Radarsat-2 fine quad pol data for developing robust technique to monitor crop condition. Similarly, polarimetric RVI and backscatter actual RVI were generated from Radarsat-2 datasets using the below Eqs. 2 and 3.

$$ {\text{Polarimetric RVI}} = \frac{{4.\min \left( {\lambda_{1} ,\lambda_{2} ,\lambda_{3} } \right)}}{{\lambda_{1} + \lambda_{2} + \lambda_{3} }} $$
(2)
$$ {\text{Actual RVI}} = \frac{{8\sigma_{HV}^{0} }}{{\sigma_{HH}^{0} + \sigma_{HV}^{0} + \sigma_{VV}^{0} }} $$
(3)

To determine the sensitive biophysical parameters, we analyze the relationship of polarimetric RVI and actual backscatter RVI with crop growth parameters of bajra crop over entire crop growth cycle (Fig. 3).

Fig. 3
figure3

Methodology flowchart for condition assessment of pearl millet crop

Results and discussion

Temporal quad polarized RVI response to various crop types

The potential of Sentinel-1 datasets for RVI analysis was compared with Radarsat-2 quad polarized data. In initial crop stage, radar backscatter consists more of soil contribution. Hence, it was assumed that both HH and VV response was same for crops only in mid to advance crop growth stages. Actual RVI was calculated using quad pol Radarsat-2 three date datasets (24th July, 17th Aug and 10th Sep, 2018) using HH-HV-VV polarizations (as both HV and VH has similar backscattering response). A truncated RVI was generated from five date Sentinel-1 datasets (2nd Aug, 14th Aug, 26th Aug, 7th Sep and 19th Sep, 2018) using VV and VH polarizations.

The RVI temporal response for various crops was plotted with error bars showing acceptable range of error. Generally, RVI values ranges from 0 to 1. Other features than vegetation (i.e. urban and water bodies) has RVI values less than 0.1. From Fig. 4a. Fodder Jowar and maize sown early (at the end of May) reaches peak vegetative phase at the start of August, shows high RVI value of 0.5 in August and after that it dips suddenly due to periodic cutting of crops for fodder purposes. The first date slight dip for Jowar and Maize was due to heavy rainfall which causes standing water in the field during image acquisition.

Fig. 4
figure4

Temporal RVI response for multiple crops: a Backscatter RVI response generated from Radarsat-2 datasets and b truncated RVI response from Sentinel-1 datasets

Paddy shows low RVI value of 0.19 than other crops during July end as the paddy fields were puddled and prepared for transplantation. RVI for paddy increased up to 0.5 in mid-September. Fallow field shows low RVI of around 0.19 in all the three dates. Desi cotton and Bt Cotton show RVI in the range of 0.2 initially increased up to 0.5 in mid of September. Bajra having close structure like maize and Jowar shows significant changes in RVI due to differences in crop calendar. Bajra, sown during the end of June shows RVI value of around 0.2 reaches 0.5 during September but Jowar and maize due to early sowing shows dip in September as shown in Fig. 4a. Because of slow growth rate of Arhar compared with other crops, almost constant RVI value of 0.3 to 0.4 was observed in all the three dates. Sesbania being the tallest crop of 5 to 6 feet in the July shows high RVI value of more than 0.6 compared to all the crops. The initial dip in the July was due to standing water in the Sesbania fields. Due to high sensitivity of RVI towards crop biomass, Sesbania having high biomass than other crop types show high RVI value. Vegetable crop like bottle gourd shows high RVI of 0.3 during peak vegetative period and again dips to 0.2 when flowering starts in September.

Temporal truncated RVI response to various crop types

Mostly truncated RVI generated from VV and VH polarizations have similar pattern as actual backscatter RVI. High temporal resolution of Sentinel -1 datasets allows better monitoring of various crop condition and vigour. Jowar and maize shows maximum RVI of 0.42 in Mid-August and then RVI values decreased due to frequent cutting of crops for fodder purposes as seen in Fig. 4b. Fallow field shows RVI values less than 0.2 in all the dates. Because of slow growth rate of Arhar crop, it shows constant RVI in the range of 0.3–0.33 in all the five dates. Bajra crop which shows RVI value 0.28 in initial stages started to increase when crop grows and shows maximum RVI value of 0.45. Sesbania shows high RVI range around 0.4–0.45 due to its high plant biomass. The initial dip observed in the first date of sesbania profile was due to standing water in the field and the dip observed in the last date was due to ploughing of sesbania field (for manure) to make it ready for another crop sowing. Vegetable crop (bottle gourd) shows low RVI in the range of 0.1–0.25. Desi cotton and Bt-cotton have RVI in the range of 0.2 in the initial date reaches maximum of 0.43 in the final date.

Temporal RVI images of radarsat-2 datasets

From above Fig. 5, brown and yellow color dominance in July indicates low RVI and most of the areas has crop with low biomass. As crop grows and reaches peak vegetative period in September, the RVI which is highly sensitive to green biomass and VWC also increases. Most of the areas which is brown and yellow in July changes to green and blue in September.

Fig. 5
figure5

Temporal backscatter RVI response in different dates generated from Radarsat-2 datasets: a 24th July, 2018, b 17th August, 2018 and c 10th September, 2018

Temporal RVI from Sentinel- 1 Datasets.

In the above Fig. 6, in first date, most of the areas were yellow and brown color with low RVI values changes to green and blue color in fourth date image. As the crop grows and reaches peak vegetative stage on fourth date (September 7), it was noticed that RVI also increases and reaches maximum on that date. But in the fifth date, due to panicle emergence in Bajra crop (as plant water content and greenness decreases), again RVI decreases and yellow shades began to appear.

Fig. 6
figure6

Temporal truncated RVI response in different dates generated from Sentinel-1 datasets: a 2nd August, 2018, b 14th August, 2018, c 26th August, 2018, d 7th September, 2018 and e 19th September, 2018

Radarsat-2 backscatter RVI relationship with various biophysical parameters of bajra crop

Crop biophysical parameters like LAI, biomass and plant height are highly variable phenomenon both spatially and temporally. RVI analysis was performed for bajra crop in relation to various biophysical parameters like wet and dry biomass, Vegetation Water Content (VWC), LAI, Plant Height, Plant volume and Plant density collected in different growth stages. RVI was highly correlated to wet biomass around R2 of 0.6, because of its higher sensitivity to moisture content in crops as depicted in Fig. 7a. No strong RVI relationship was observed with plant dry biomass as R2 was found to be less than 0.5 Fig. 7b. From Fig. 7a RVI increased with increase in crop biomass up to 5 kgm−2 and beyond that saturation occurs in C-band. To overcome this issue, it is proposed to use long wavelength L or S band for higher biomass crops.

Fig. 7
figure7

Backscatter RVI relationship with various biophysical parameters: aWet Biomass vs RVI (y = -5E-09x2 + 6E-05x + 0.2307), b Dry Biomass vs RVI (y = -8E-08x2 + 0.0002x + 0.2512), c LAI vs RVI (y = 0.0548x + 0.2122), d VWC vs RVI (y = -6E-09x2 + 7E-05x + 0.2327), (e) Wet Biomass vs VWC (y = 0.7667x—5.5255), (f) Plant Height vs RVI (y = -5E-07x2 + 0.001x + 0.2157), g Plant Volume vs RVI (y = -1E-07x2 + 0.0004x + 0.2166) and h Plant Density vs RVI (y = 0.0136x + 0.1819)

LAI shows linear relationship with RVI of R2 around 0.57. Unlike NDVI, there is no saturation observed in RVI with LAI in different growth phases Fig. 7c. RVI provides strong polynomial relationship with VWC of around R2 0.6 till 4 kgm−2 and then saturation was observed as shown in Fig. 7d. Plant volume for collected samples were found using Archimedes principle and achieved R2 of around 0.52 Fig. 7g. Plant height is one of the important parameters to indicate the growth stages of crops. RVI shows linear trend with plant height (R2 = 0.47). Plant density provides R2 around 0.66 with RVI which proved to have significant correlation. When both ground-measured Wet Biomass and VWC was plotted, close linear relationship was observed which has high R2 of 0.95, as both parameters are highly dependent on each other Fig. 7e.

Crop age in relation with RVI & crop parameters

Crop age is an important parameter for determining sowing date of particular crop especially in uneven sowing regions. An attempt was made to establish relationship between crop age collected from ground truth survey to quad polarized Radarsat-2 RVI as depicted in Fig. 8a. Crop age was also correlated with other ground based biophysical parameters.

Fig. 8
figure8

Relationship of crop age with various biophysical parameters of pearl millet crop: aAge vs RVI (y = -6E-06x2 + 0.0028x + 0.1828), b Age vs Wet biomass (y = -0.4212x2 + 111.05x—1949.8), c Age vs VWC (y = -0.2623x2 + 74.51x – 1186), and d Age vs Plant height (y = -0.0155x2 + 4.0726x—48.481)

When crop age was plotted against RVI values, strong polynomial relationship was observed with an R2 in the range of 0.6 as seen in Fig. 8a. Crop age shows high correlation with plant height with R2 more than 0.7 as observed in Fig. 8d. It was observed that RVI value shows an increasing trend with crop height up to 120 cm, beyond that it saturates in C-band. Wet Biomass and VWC shows polynomial relationship with crop age having R2 in the range of 0.6 as seen in Fig. 8b, c.

Truncated RVI Relationship with various Biophysical parameters of Bajra crop

An attempt was made to establish relationship of crop biophysical parameters with truncated RVI to assess the condition of bajra crop. The results were found to be promising in Sentinel-1 datasets for crop monitoring and vigour assessment. The wet biomass shows R2 more than 0.5 with RVI obtained using VV-VH polarizations as shown in Fig. 9a. As like actual RVI obtained from Radrasat-2 datasets, here also RVI have increasing trend up to 4 kgm−2 and beyond that saturation found in C-band.

Fig. 9
figure9

Truncated RVI relationship with various biophysical parameters: a Wet Biomass vs RVI (y = -2E-09x2 + 3E-05x + 0.2004), b Dry Biomass vs RVI (y = -6E-08x2 + 0.0002x + 0.2125), c LAI vs RVI (y = 0.05x + 0.1739), d VWC vs RVI (y = -2E-09x2 + 3E-05x + 0.2146), e Wet Biomass vs VWC (y = 0.7667x—5.5255), f Plant Height vs RVI (y = -5E-06x2 + 0.002x + 0.1315), g Plant Volume vs RVI (y = -1E-07x2 + 0.0003x + 0.2126) and h Plant Density vs RVI (y = 0.0095x + 0.201)

Dry biomass does not show strong correlation with RVI, where R2 is around 0.4 as depicted in Fig. 9b. From Fig. 9c, it was observed RVI increases with increase in LAI showing linear relationship and no saturation was observed in advanced growth stage. Crop parameters like plant height and volume shows strong polynomial relationship with RVI in wide dynamic range having R2 more than 0.5 as seen in Fig. 9f, g. Plant density shows linear relationship with RVI as observed in Fig. 9h. Wet biomass and VWC content have close linear relationship with R2 in the range of 0.95 as shown in Fig. 9e.

Crop age calculated during five satellite passes was plotted against the truncated RVI. The high polynomial relationship was observed with R2 more than 0.55 as shown in Fig. 10. Other ground measured parameters against crop age gives similar relationship as obtained by Radarsat- 2 plots.

Fig. 10
figure10

Plant Age vs truncated RVI (y = -2E-05x2 + 0.0039x + 0.1376) extracted from Sentinel-1 VV-VH polarizations

Polarimetric RVI relationship with crop growth variables

Polarimetric RVI was calculated using single look complex quad polarimetric Radarsat-2 datasets containing both phase and amplitude information. The sensitivity of polarimetric RVI towards various biophysical parameters of bajra crop were tested. Polarimetric RVI shows higher saturation level for all the crop variables when compared with truncated and actual RVI. Here RVI increases linearly as crop grows till 6 kgm−2 with R2 above 0.7 and beyond that it starts to saturate as shown in Fig. 11a. The similar linear trend is observed with dry biomass showing an R2 of 0.6 Fig. 11b. RVI having higher sensitivity towards VWC shows polynomial relationship with R2 above 0.7 Fig. 11c. Polarimetric RVI increases with increase in LAI even in advanced crop growth stages as indicated in Fig. 11d, as NDVI becomes saturated after a level of growth in crop. RVI shows linear trend with plant density and plant height with R2 of 0.75 and 0.6 as seen in Fig. 11f, g. Polarimetric RVI shows higher sensitivity towards plant volume with R2 above 0.7 (Fig. 11e). Crop age shows good polynomial trend with polarimetric RVI proving its potential in finding showing date of crop.

Fig. 11
figure11

Polarimetric RVI relationship with various biophysical parameters: aWet Biomass vs RVI (y = 1E-10x2 + 4E-05x + 0.3537), b Dry Biomass vs RVI (y = -2E-08x2 + 0.0002x + 0.3771), c VWC vs RVI (y = -3E-09x2 + 7E-05x + 0.3538), d LAI vs RVI (y = 0.0506x + 0.3698), e Plant Volume vs RVI (y = -2E-07x2 + 0.0004x + 0.3751), f Plant Density vs RVI (y = 0.0188x + 0.3407), g Plant Height vs RVI (y = -3E-06x2 + 0.002x + 0.3017), and h Plant age vs RVI (y = -2E-05x2 + 0.0059x + 0.2881)

Conclusion

The present study brought out the direct sensitivity of SAR towards Vegetative Water Content of crop. This will help in regular monitoring of water stress occurring in crops. The study shows the applicability of RVI for monitoring crop growth in different stages and assessing vigour of the crop. Truncated RVI from Sentinel-1 datasets has potential in monitoring crop condition in regional basis due to the wider swath coverage. Deploying high temporal repetivity of Sentinel-1 datasets helps to observe the finer RVI response and changes occurring on different crops in short span of time. RVI have the prospect of using in soil moisture models in vegetated fields as vegetation mask. Moreover, RVI can also be utilized in place of NDVI for vegetation monitoring in cloudy kharif season. For higher biomass crops during advanced growth stages, Polarimetric RVI obtained from information rich Radarsat-2 was found to be promising as HH polarizations captures higher dynamic range. This study also reveals the need of using L, S or P- band in case of higher biomass crop like pearl millet as beyond 6 kgm−2 saturation occurs in C-band. The upcoming NISAR mission gives a chance to address the saturation issue faced in higher biomass crops. The strong relationship observed between RVI and crop age will be used in predicting sowing date. The promising results shows the further applicability of this robust technique to other crops.

References

  1. 1.

    Shanmugapriya, S., Haldar, D., & Danodia, A. (2019). Optimal Datasets Suitability for Pearl Millet (Bajra) Discrimination using Multiparametric SAR Data. Geocarto International, 6049, 1–14. https://doi.org/10.1080/10106049.2019.1585485.

    Article  Google Scholar 

  2. 2.

    Bouvet, A., Le Toan, T., & Lam-Dao, N. (2009). Monitoring of the Rice Cropping System in the Mekong Delta Using ENVISAT/ASAR Dual Polarization Data. IEEE Transactions on Geoscience and Remote Sensing, 47(2), 517–526. https://doi.org/10.1109/TGRS.2008.2007963.

    Article  Google Scholar 

  3. 3.

    Dave, V. A., Haldar, D., Dave, R., Misra, A., & Pandey, V. (2017). Cotton crop biophysical parameter study using hybrid/compact polarimetric RISAT-1 SAR data. Progress in Electromagnetics Research, 57, 185–196. https://doi.org/10.2528/PIERM16121903.

    Article  Google Scholar 

  4. 4.

    McNairn, H., Shang, J., Jiao, X., & Champagne, C. (2009). The Contribution of ALOS PALSAR Multipolarization and Polarimetric Data to Crop Classification. IEEE Transactions on Geoscience and Remote Sensing, 47(12), 3981–3992. https://doi.org/10.1109/TGRS.2009.2026052.

    Article  Google Scholar 

  5. 5.

    Shanmugapriya, S., Haldar, D., & Danodia, A. (2019). Time series Sentinel-1A profile analysis for heterogeneous Kharif crops discrimination in North India. URSI AP-RASC 2019, New Delhi, India, 09–15 March 2019.

  6. 6.

    Jin, Y.-Q., & Liu, C. (1997). Biomass retrieval from high-dimensional active/passive remote sensing data by using artificial neural networks. International Journal of Remote Sensing, 18(4), 971–979. https://doi.org/10.1080/014311697218863.

    Article  Google Scholar 

  7. 7.

    Kim, Y., Jackson, T., Bindlish, R., Lee, H., & Hong, S. (2013). Monitoring soybean growth using L-, C-, and X-band scatterometer data. International Journal of Remote Sensing, 34(11), 4069–4082. https://doi.org/10.1080/01431161.2013.772309.

    Article  Google Scholar 

  8. 8.

    Tucker, C. J. (1979). Red and photographic infrared linear combinations for monitoring vegetation. Remote Sensing of Environment, 8(2), 127–150. https://doi.org/10.1016/0034-4257(79)90013-0.

    Article  Google Scholar 

  9. 9.

    Jackson, T. J., & Schmugge, T. J. (1991). Vegetation effects on the microwave emission of soils. Remote Sensing of Environment, 36(3), 203–212. https://doi.org/10.1016/0034-4257(91)90057-D.

    Article  Google Scholar 

  10. 10.

    Gao, B.-C., & Goetzt, A. F. H. (1995). Retrieval of equivalent water thickness and information related to biochemical components of vegetation canopies from AVIRIS data. Remote Sensing of Environment, 52(3), 155–162. https://doi.org/10.1016/0034-4257(95)00039-4.

    Article  Google Scholar 

  11. 11.

    Kim, Y., & van Zyl, J. (2001). Comparison of forest parameter estimation techniques using SAR data. IEEE 2001 International Geoscience and Remote Sensing Symposium, 3, 1395–1397. https://doi.org/10.1109/IGARSS.2001.976856.

    Article  Google Scholar 

  12. 12.

    Kim, Y., Jackson, T., Bindlish, R., Lee, H., & Hong, S. (2012). Radar vegetation indices for estimating the vegetation water content of rice and soybean. IEEE Geoscience and Remote Sensing Letters, 9(4), 564–568. https://doi.org/10.1109/LGRS.2011.2174772.

    Article  Google Scholar 

  13. 13.

    Tugrul Yilmaz, M., Hunt, E. R., Goins, L. D., Ustin, S. L., Vanderbilt, V. C., & Jackson, T. J. (2008). Vegetation water content during SMEX04 from ground data and landsat 5 Thematic mapper imagery. Remote Sensing of Environment, 112(2), 350–362.

    Article  Google Scholar 

  14. 14.

    Kim, Y., Jackson, T., Bindlish, R., Hong, S., Jung, G., & Lee, K. (2014). Retrieval of wheat growth parameters with radar vegetation indices. IEEE Geoscience and Remote Sensing Letters, 11(4), 808–812. https://doi.org/10.1109/LGRS.2013.2279255.

    Article  Google Scholar 

  15. 15.

    S. Mohan, A. Das, D. Haldar, and S. Maity (2014) Monitoring and Retrieval of Vegetation Parameter using Multi-frequency Polarimetric SAR Data. Asia-Pecific Conf. Synth. Aperture Radar 330–333

  16. 16.

    Haldar, D., Dave, V., Misra, A., & Bhattacharya, B. (2018). Radar vegetation index for assessing cotton crop condition using RISAT-1 data radar vegetation Index for assessing cotton crop. Geocarto International. https://doi.org/10.1080/10106049.2018.1516249.

    Article  Google Scholar 

  17. 17.

    Ferrazzoli, P., Paloscia, S., Pampaloni, P., Schiavon, G., Solimini, D., & Coppo, P. (1992). Sensitivity of microwave measurements to vegetation biomass and soil moisture content: a case study. IEEE Transactions on Geoscience and Remote Sensing, 30(4), 750–756. https://doi.org/10.1109/36.158869.

    Article  Google Scholar 

  18. 18.

    Brisco, B., Brown, R. J., Gairns, J. G., & Snider, B. (1992). Temporal ground-based scatterometer observations of crops in western canada. Canadian Journal of Remote Sensing, 18(1), 14–21. https://doi.org/10.1080/07038992.1992.10855138.

    Article  Google Scholar 

  19. 19.

    Chauhan, S., Srivastava, H. S., & Patel, P. (2018). Wheat crop biophysical parameters retrieval using hybrid-polarized RISAT-1 SAR data. Remote Sensing of Environment, 216, 28–43. https://doi.org/10.1016/J.RSE.2018.06.014.

    Article  Google Scholar 

  20. 20.

    Saxena, N., & Rathore, N. (2013). A review on speckle noise filtering techniques for SAR images. International Journal of Advanced Research in Computer Science and Electronics Engineering , 2(2), 2277–9043.

    Google Scholar 

  21. 21.

    Srivastava P. K., O’Neill P., Cosh M., Lang R., Joseph A. (2015) Evaluation of Radar Vegetation Indices for Vegetation Water Content Estimation Using Data from a Ground-Based SMAP Simulator. In 2015 IEEE International Geoscience and Remote Sensing Symposium (IGARSS) 1296–1299. IEEE.

  22. 22.

    Haldar, D., Chakraborty, M., Manjunath, K. R., & Parihar, J. S. (2014). Role of polarimetric SAR data for discrimination/biophysical parameters of crops based on canopy architecture. International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. https://doi.org/10.5194/isprsarchives-XL-8-737-2014.

    Article  Google Scholar 

Download references

Acknowledgements

We would like to thank European Space Agency (ESA) for providing Sentinel-1 datasets and SNAP software for RVI analysis. The research was done under SUFALAM project and we are also thankful to Indian Institute of Remote Sensing (IIRS) for providing necessary resources for carrying out this research.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Shanmugapriya Selvaraj.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Selvaraj, S., Haldar, D. & Srivastava, H.S. Condition assessment of pearl millet/ bajra crop in different vigour zones using Radar Vegetation Index. Spat. Inf. Res. (2021). https://doi.org/10.1007/s41324-021-00380-y

Download citation

Keywords

  • Radar vegetation index
  • Biophysical parameters
  • Vegetation water content
  • Wet biomass
  • Leaf area index
  • Plant volume
  • Polarimetric RVI
  • Backscatter RVI and truncated RVI