Back to Course

Canadian SAR Mini-MOOC: Winter-Water-Warming

0% Complete
0/0 Steps
Lesson 1, Topic 1
In Progress

Data Collection and Processing

This topic covers the important aspect of how we combine satellite- and ground-based observations to better map Canada’s agricultural resources. Here, Thierry Fisette explains how AAFC collects, processes and integrates satellite imagery and field data to create Canada’s Annual Crop Inventory.

Input Data: Ground Reference Information

Ground reference, or in situ, information is critically important for training and validating numerical models and for obtaining successful crop classifications. Generally speaking, crop insurance data sets are the most accurate, detailed, and complete sources of geospatial crop information in Canada. Currently, annual crop insurance data are provided by four Canadian Provinces: Alberta, Saskatchewan, Manitoba and Québec. Together they represent 87 percent of Canada’s agricultural area.

caption, credit / Source ??

For those provinces where crop insurance data cannot be accessed, AAFC staff is collecting ground-truth information by way of in-situ observations. In 2019, more than 70,000 of these observations were gathered in Canada; 70 percent of them were used to train the classifier, and the remaining 30 percent were used to validate the final map product.

Here comes the real GIF!!

Field Data Collection

The collection of field data is an integral part of AAFC’s production of the national Annual Crop Inventory. The following video illustrates the process and procedures.


By loading the video, you agree to YouTube’s privacy policy.
Learn more

Load video

The actual process of AAFC personnel logging pertinent crop inventory information from extensive road-side observations is illustrated in the following video.


By loading the video, you agree to YouTube’s privacy policy.
Learn more

Load video

Input Data: Optical and Radar Imagery

Optical Landsat-8 and Sentinel-2 satellite data can be quite adequate to classify crops, provided that sufficient cloud-free data is available during critical periods of the growing season. Classification accuracies of more than 85 per cent have been achieved. However, they can be significantly degraded by data collection gaps.

Our research and operational experience have shown that the combined use of optical and SAR satellite data can increase overall accuracies by up to 16 per cent. Both data streams provide unique and valuable information. Optical imagery is extremely useful for differentiating crop types, whereas RADARSAT SAR data has proven valuable for determining plant structure and for filling gaps in the optical image record.

Prior to using optical and radar data in crop classification procedures used and crop inventory generation, some pre-processing steps must be carried out. All images are orthorectified using a 3-D multi-sensor physical model. Bottom of Atmosphere (BOA) reflectance images are derived from optical data. The 2020 crop inventory called for the collection of RADARSAT Constellation Mission (RCM) data to extract compact polarimetry variables for crop identification.

More than 1500 RCM images, or one complete national acquisition per month for five months, are required for the SAR classification process, which also includes radiometric calibration and speckle noise deduction. For both radar and optical data sets, along-track images of the same date are mosaicked together to form large classification regions and include more training sites.

(overlaid Optical & SAR images)  Caption ?   Credit:  ???

Decision Tree Classification

Since 2009, the AAFC has created its Annual Crop Inventory (ACI) by applying the Decision Tree (DT) algorithm to various combinations of optical and SAR imagery using discriminate functions estimated empirically from hundreds of thousands of ground-based (in situ) training data samples. The AAFC operations use the C5.0 Decision Tree (DT) method for a variety of reasons. They include the ability to handle discrete data, processing speed, independence of the distribution of class signatures, interpretable classification rules, cost-effectiveness and accuracy.

AAFC performs the annual DT classifications region-by-region due to dynamic nature of crop rotations, crop growth stages, and harvest patterns. While each classification region combines several dates of optical and SAR imagery, the actual combination of imagery per region can vary based on data availability.

Accuracy of Annual Crop Inventory

The ACI consistently achieves an overall accuracy of 85 per cent at the national scale. Nonetheless, actual mapping accuracy varies from crop-to-crop, region-to-region, and year-to-year, depending on satellite data availability and in-situ training data. In general, the highest mapping accuracies of more than 90 per cent are achieved where crops display significantly different spectral characteristics at the time of satellite data acquisition, as is often the case for the Canadian Prairie Provinces of Saskatchewan and Alberta (see bar graph “2011-2019 Per-Province Crop Overall Accuracy”). Elsewhere, accuracies may range from approximately 75 to 85 per cent. Variability relates to limitations associated with the in-situ and satellite data used in the classification process.

First, there can be significant Province-to-Province variability in the number, density and quality (i.e. detail and accuracy) of in-situ data used for training the Decision Tree (DT) classification and its validation. Second, there is often significant regional variability in cloud cover, which can limit the amount of optical satellite data (currently Sentinel-2 and Landsat-8) available for input to the DT classification.

For example, while as many as 3 or 4 clear-sky optical images can be acquired over locales within the Canadian Prairie Provinces during each growing season, this availability can be reduced to a single image for some locales in Canada’s coastal regions. This issue should become less problematic in the future as access to new streams of optical and radar imagery increases.


Final Maps Distribution

The annual crop inventory datasets are publicly available at no cost at as part of the Canadian federal government commitment to open data. The crop inventory mosaics are distributed in raster format (GeoTIFF) in the Albers Equal Area projection. The AAFC metadata – information that accompanies and describes the dataset – are based on ISO 19131 standards. Bilingual (French and English) metadata are available. Annual crop maps are typically delivered approximately six months following the end of growing season.


Further references and links related to this sections are located here.

Next Lesson:

Lake and River Ice

With this topic we end the lesson about mapping and monitoring of Canada’s agricultural landscapes. Let’s move on to the next lesson to learn about how Canada is using SAR to monitor lake and river ice.

Please complete the quiz below to advance to the next lesson.

Click the ‘Submit’ Button to rate this topic.