Horticultural Crop Assessment using Satellite Data (Coordinated Horticulture Assessment and Management using geo-informatics: CHAMAN) S. Mamatha 1, B. K. Bhattacharya 2, Uday Raj 3, H P Sharma 4, B K Handique 5, Mamta Saxena 6 & S S Ray 1 Email: mamatha.ncfc@nic.in 1 Mahalanobis National Crop Forecast Centre. DAC&FW, New Delhi 2 Space Applications Centre, ISRO, Ahmedabad, 3 Natinal Remote Sensing Centre and RRSCs,ISRO, Bengaluru 4 National Horticulture development Foundation, Nasik 5 North eastern Space Application Centre,ISRO, Shillong 6 Department of Agriculture, cooperation and Farmers welfare
Objectives Overview of CHAMAN Project Area assessment and production forecasting of 7 major horticultural crops in selected districts of major states. Geospatial Applications for Horticultural Development and Management Planning. Detailed scientific field level studies for developing technology for crop identification, yield modelling and disease assessment
Objective 1. Crop Inventory Crop Crop States type Fruit Banana AP, Gujarat, Karnataka, Maharashtra, TN Mango Citrus Vegetables Potato AP, Bihar, Gujarat, Karnataka, Telangana, UP AP, Gujarat, Punjab, Maharashtra, MP, Telangana Bihar, Gujarat, Punjab, UP, WB Onion Gujarat, Karnataka, Maharashtra, MP Tomato AP, Bihar, Karnataka, MP, Odisha, Telangana, WB Spices Chilli AP, Karnataka, MP, Odisha, Telangana, WB Total 185 Districts Only major crop growing districts of these states are considered
Satellite Data Used for Crop Inventory Satellite Sensor Resoluti Spectral Sets of data Crops on Bands used required at a time Resourcesat AWIFS 56m NIR and Red Single date/ Multi-Date Potato. -2 56m NDVI product Fortnightly Potato LISSIII 23.5m NIR, Red and Single date/ Multi-Date Potato, Onion, Banana Green LISSIV 5.8m NIR, Red and Green Single date/ Multi-Date Onion, Chili, Tomato, Mango, Citrus IRS-P5 Cartosat 1 2.5m PAN Single date Mango, Citrus Landsat 8 OLI 30m SWIR, NIR, Red, Single date/ Multi-Date Potato, Onion Green and Blue Sentinel-2A MSI 10m NIR, Red, Green and Blue Single date/ Multi-Date Potato, Onion, Chili, Tomato
General Methodology Pre Processing of satellite data Ground truth data collection Satellite Image Classification Post-classification analysis Quality evaluation and Accuracy assessment Area Estimation Map Preparation Bhuvan Interface Classification Characteristics Classifiers Crop Techniques Pixel based techniques Object Based Techniques Each pixel is assumed pure and typically labeled as a single land use land cover type Geographical objects, instead of individual pixels are compared to the basic unit Unsupervised (e.g. K- means, ISODATA); Supervised (e.g. Maximum Likelihood- MXL) Image segmentation an object based image analysis techniques (e.g. E-cognition, ArcGIS/ Imagine Objective) Onion, Chili, Potato and Banana Mango, Citrus and Banana
Potato Classification using AWiFS & Sentinel Data
Onion Classification using Sentinel Data
Tomato and Chilli Classification using LISS III/Sentinel/Landsat/LISS IV Data
Orchard Mapping using LISS IV/ LISS IV + Cartosat Data Citrus orchards in LISS IV data in Hoshiyarpur, Punjab Citrus orchards in LISS IV +Cartosat merged data in Bijapur, Karnataka Mango orchards in LISS IV + Cartosat merged data in Saharanpur District, UP Mango Orchards in LISS IV + Cartosat merged data in Sitapur, UP
MANGO ORCHARD Sitapur district of Uttar Pradesh 99_52_B_15 April 2013, LISS 4 FCC with Ground Truth points Classified Mango orchards overlaid on LISS 4 FCC with Ground Truth points IRS R-2 LISS-IV, Pixel based Classified output Refined Classified Output by manual editing over BHUVAN image as a Basemap IRS R-2 LISS-IV, Object based Classified output Classification Method Derived area (ha) Pixel based 12117 Object based 15387 BHUVAN Updated 15440 *Reference area (ha) 15389 Final Mango Orchards mapped for entire Sitapur district derived Mango Orchard area= 15440 ha
CHAMAN in BHUVAN Geoportal http://bhuvan-staging.nrsc.gov.in/projects/moa_chaman/
Mango Orchard of Sitapur District, Bhuvan-CHAMAN Portal Scale: 1:15000 Scale: 1:5000
Deviation Analysis of Crop Statistics Crops MBE ('000 ha) RMSE ('000 ha) IA (unit less) ME (unit less) r 2 Number of data points Citrus -2.64 4.64 0.92 0.76 0.87 9 Mango -2.10 4.89 0.97 0.90 0.94 26 Banana -3.56 7.26 0.84 0.61 0.82 9 Onion -14.98 25.75 0.84 0.52 0.75 15 Potato -0.06 6.47 0.98 0.91 0.91 44
Horticultural Development using Geospatial Technology S.N. Component Activity 1 Site Suitability Introduction/expansion of Horticulture development activities in North Eastern States (One district in each state) 2 Post-Harvest Infrastructure Assessing the potential for new cold storage sites for in Bihar and UP State 3 Crop Intensification Understanding the scope of improving crop intensity through horticulture in selected districts of Haryana and Madhya Pradesh 4 GIS database creation Characterization of orchards and GIS database creation of various layers and uploading on Bhuvan platform 5 Orchard Rejuvenation Identifying plantations /orchards that needs Rejuvenation in one District in UP and One district of Karnataka/Gujarat/WB using remote Sensing data 6 Aqua-horticulture Developing plans for promoting aqua-horticulture in 1 2 districts in Bihar and Odisha state
Site Suitability Analysis for Horticulture Expansion in NER -States Land Suitability Analysis for Mango Plantation in Nuzvid mandal, AP Perspective view of Jhum land clusters Sl No State District Crop 1 Arunachal Papumpare Orange Pradesh 2 Assam Goalpara Banana 3 Manipur Senapati Pineapple 4 Meghalaya Jaintia Hills Turmeric 5 Mizoram Champhai Grape 6 Nagaland Dimapur Pineapple 7 Sikkim West Sikkim Cardamom 8 Tripura West Tripura Banana
Geospatial Applications: Post-Harvest Infrastructure Existing Proposed
Geospatial Applications: Aqua-Horticulture Legend Roads Railways!( Priority1 Villages # Priority2 Villages Major Towns District Boundary Potential Wetlands Priority 2 Potential Wetlands Priority 1 Makhana Ponds Not Potential Zones Category Area (Ha) Number Area under Foxnut cultivation 339.6 109 Ponds under priority 1 99.0 47 Ponds under priority 2 102.9 33
Geospatial Applications: Orchard Rejuvenation Potential Area Identification Based on NDVI Change Based on Orchard Structure
Geospatial Applications: Crop Intensification, Bhiwani, Haryana Full Seasonal Fallow 2015 Fallow Period 2015 Number of months Area(ha) 2015 Rabi 2015 Kharif 2015 Summer 1 month 82231 134075 76850 2 month 11412 36543 234219 3 month 1637 6231 86425 4 month 24700 Legend Annual Fallow Built_up Water Taluk Boundary
Precision Farming Study Study Area: Karsul Village, Niphad Taluk, Nasik Data Collected: Ground Spectral, UAV, High Resolution Satellite, Crop and Soil Parameters Activity: Phenology Mapping, Variability Assessment, Crop and Soil Parameter Retrieval
180 Normalized Difference Vegetation Index (NDVI) 170 160 150 140 130 120 Temporal NDVI Profiles of Grapes with Varying Vigour Grape-Nov Fruit Pruning Low Vigour Grape-Nov Fruit pruning High Vigour Figure 5: Temporal spectral profile of grapes with varying vigour identified by stacking monthly NDVI Landsat imagery of 2013-14, 2014-15 and 2015-16 for Karsul village, Niphad block, Nasik district Dates 180 170 Normalized Difference Vegetation Index (NDVI) 160 150 140 130 120 Temporal NDVI Profiles of Grapes with Varying Grape Young Orchard Phenology Grape-Oct Fruit pruning High Vigour Sept Fruit Pruning Oct Fruit Pruning Dates Nov Fruit Pruning Grape-Sept Fruit Pruning Low Vigour Grape-Nov Fruit pruning High Vigour Nov Fruit Oct Fruit Pruning Sept Fruit Pruning Pruning Figure 7: Temporal spectral profile of grapes with varying phenology identified by stacking monthly NDVI Landsat imagery of 2013-14, 2014-15 and 2015-16 for Karsul village, Niphad block, Nasik district
Summary For crops like Potato, Mango, Citrus and Banana use of remote sensing data for assessment have been feasible, but for other crops accuracy is still an issue, due to scattered and small fields, mixed cropping, multiple seasons and short duration. Yield estimation for horticultural crops, especially fruit crops, is a problem due to multiple picking. However use of satellite data and geospatial tools has shown a great promise for horticultural development, especially for infrastructure development and horticultural expansion.
Acknowledgment Indian Space Research Organization (SAC, NRSC, NESAC) Department of Agriculture, Cooperation & Farmers Welfare (Hort Div) National Horticultural Research & Development Foundation India Meteorological Department State Horticulture Departments ICAR:NRCG Team Members 1. M M Kimothi 2. Seema Sehgal 3. Shreya Roy 4. Aditi Srivastava 5. Niti Singh 6. Gargi Upadhyay 7. Moreshwar Karale Thank you