The application of Remote Sensing Technology for the identification of sowing varieties for a
selected area using the Azersky satellite Imageries.
The main purpose of crop yield is distinguish agricultural areas, allocation and different
plant areas based on Azersky satellite images on selected plant species (cereals, cotton, tobacco,
patch and sugar beet, etc).
Identification and mapping of crops is of utmost importance for a number of reasons. Key
activities includes, identifying the crop types and delineating their extent. Remote sensing offers
an efficient and reliable means of collecting the information required, in order to map crop type
and acreage.
A human analyst attempting to classify features in an image uses the elements of visual
interpretation to identify homogeneous groups of pixels which represent various features or land
cover classes of interest. Digital image classification uses the spectral information represented by
the digital numbers in one or more spectral bands and attempts to classify each individual pixel
based on this spectral information. This type of classification is termed spectral pattern recognition.
In either case, the objective is to assign all pixels in the image to particular classes or themes (e.g.
water, coniferous forest, deciduous forest, corn, wheat, etc.). The resulting classified image is
comprised of a mosaic of pixels, each of which belong to a particular theme, and is essentially a
thematic "map" of the original image.
Spatial resolution refers to the discernible detail in the image. Detailed mapping of
wetlands requires far finer spatial resolution than does the regional mapping of physiographic
areas.
Satellite images are used as mapping tools to classify crops, examine their health and
viability and monitor farming practices. Agricultural applications of remote sensing are crop type
classification, crop yield estimation, crop condition assessment, mapping of soil characteristics
and mapping of soil management practices. Identifying and mapping crops is important for a
number of reasons. Maps of crop type are created by national and international agricultural and
insurance agencies. This serves the purpose of forecasting grain supplies, collecting crop
production statistics, facilitating crop rotation records, mapping soil productivity, identification of
factors influencing crop stress, assessment of crop damage due to storms and drought, and
monitoring farming activity.
Agricultural cropland mapping using 6 times across the country for multicultural
landscapes and pansharp for crop species.
Multispectral images obtained from the Azersky satellite has vegetation sensitive spectral
bands and its 6 meter spatial resolution which facilitates the detection of agricultural field
boundaries and also used to understand how to plant seeds at early stages of development and to
describe seasonal crop varieties.
The following stages of implementation are described in the guide:
Image acquisition;
Orthorectification;
Field reconnaissance;
Preparation of plant cover and planting maps.
Accuracy assessment.
Based on practice, 5% of the total polygons should be obtained from all types of field
samples. Most examples include wheat, cotton, tobacco, barley, grapes, peas, and clover. There
are also spring sowing, nuts, cotton, bean, sunflower and grain samples. Crop identification and
mapping benefit from the use of multitemporal imagery to facilitate classification by taking into
account changes in reflectance as a function of plant phenology (stage of growth). This in turn
requires calibrated sensors, and frequent repeat imaging throughout the growing season. For
example, crops type wheat and cotton may be easier to identify when they are flowering. Cause
both of these during month to month the spectral reflectance change and the timing of the
flowering.
1624 field samples were collected from 15 different classes. Field samples separate
“training” and “validation” groups. A random 70–30% splitting of the 1624 were used to separate
1137 samples for “training” and the rest 487 for “validation”. The 1624 training samples were used
to create knowledge through ideal spectral libraries. Ground data samples repository collected
during the field visit includes mostly pure classes.
The Azersky satellites (Spot-7) provide images every 26 days with a spatial resolution of
1.5 m. In total, 24 satellite images used for classification from February till end of July 2019.
Including 8 different time satellite images (4 * 6 = 24) with each 4 bands (24 bands with 6 different
descriptions for summer sowing).
Maximum Likelihood Classification (MLC) is the most popular classification method used
in Remote Sensing. The main purpose here is that the pixel at the intersection is calculated and
which is included in the class that receives a high probability value from the pixel. The images are
combined (mosaic) and ultimately to illustrate multitemporal images.
Unsupervised classification are based on the software analysis of an image and computer
uses techniques to determine which pixels are related and groups them into classes. The user can
specify which algorism the software will use and the desired number of output classes. But spectral
properties of informational classes change over time so user can’t always use same class statistics
when moving from one image to another.
To validate the classification results, the resulting crop type is compared with the actual
crop type of the validation data. Of the 1624 samples collected, 487 (30%) were used for quality
control purposes. With the sampling within the sample areas, 3289 checkpoints were created using
a random method and a qualitative calculation was performed using the Confusion matrix method.
Quality accuracy consists of analysis and verification of field samples as a result of the
classification. Using the confusion matrix (or error matrix), P (produce) and U (User) accuracy are
calculated. Comparing the image with the results of its interpretation, we can see errors and
roughly estimate their size. But if we need a reliable accuracy assessment, we can’t do without
quantitative methods of evaluation.
Result shows that the crop maps derived using seasonal features achieved an overall
accuracy of more than 87%. To explore in detail which crop types are classified correctly or
incorrectly and which confusions between crop types may occur, the final classification results
end of July are further analysed using a confusion matrix. The rows of the confusion matrix
represent the classification result, whereas its columns represent the validation data. The
producer’s accuracy (P) indicates how many fields are assigned to the correct crop type and the
user’s accuracy (U) represents the probability that a field belongs to its assigned crop type. If the
result is not satisfactory, re-classification is done. If there is close to 100% n, P and U get the
accuracy of the value: When there are only a few field samples on the class, the sample patterns
are very small in size and the images that fit exactly the vegetation period are selected. This study
shows the potential of early-season crop type mapping, which is useful for crop management.