Aerial NDVI Data Collection For Use in Surface Algae Analysis

Katie Gilmore, PAAP Drones

Abstract:  

The Normalized Difference Vegetation Index (NDVI) is widely used to measure and monitor plant health for agriculture and land resource management.  With the right technology, plant matter on water surfaces can also be evaluated using NDVI.  Using drones to calculate high resolution NDVI data will enable water management personnel to quickly identify patterns of algae in bodies of water, and can help lead to successful mitigation efforts.

 

Introduction:  

In a state with more than 21,000 lakes, evaluating Minnesota’s water bodies can be a time consuming and expensive task.  Monitoring surface water health and potential nutrient runoff is important as construction, development, and commercial land-use increases around our bodies of water.  The ability to efficiently identify sources of nutrient runoff can be a helpful first step in algae control.  Traditionally, water sampling has been a time-consuming and costly method to evaluate pond health and determine sources of excess nutrient supply.  The ability to use aerial data collection to identify areas of “healthy” algae (areas where algae is prevalent and growth is productive) is an important tool that is available to water resource management teams.

NDVI is calculated by measuring the reflection of visible light and near infrared (nir) light, and is an indication of chlorophyll activity (which is prevalent in healthy, leafy vegetation).  Healthy vegetation absorbs visible light, reflects near infrared light, and returns positive NDVI values between .1 to 1, with positive, higher values indicating healthier vegetation.   Objects that do not contain chlorophyll activity, such as water, rocks, dirt, and urban structures, result in negative NDVI values.   

 

Methodology:

A DJI Inspire 1 was flown over water body LP47 in Eagan, MN on October 11th at 2pm, at approximately 340ft above ground level.  The drone carried a Sentera Multispectral Sensor, capturing 59 images each of red, green, blue (RGB) and nir imagery.  Imagery post processing was completed in Pix4D utilizing the Index Calculator module, and further analyzed in ArcMap.  The following equation was used to calculate NDVI, normalized to a scale from -1 to 1.  

Typical NDVI values are displayed in a 5-class gradient from Red to Green, where Red trends to -1 and Green trends to +1.  Water and other non-vegetative areas with negative NDVI values and are displayed in red, in contrast to orange, yellow, and green vegetation that represent increasingly positive values.  In the case of bodies of water that contain surface algae, readouts can vary significantly, and areas containing surface algae will result in positive NDVI values.

Figure 1 shows the aerial view of LP47 in RGB.  Surface algae is visible on the north-east and west areas of the pond.  There is significant tree and brush growth around the shoreline and the lake is bordered by a road to the east, a parking lot to the north, and a residential area to the south.

 

RGB Aerial Imagery

Figure 1: LP47 RGB

 

Figure 2 provides an informative view of surface NDVI values for LP47.  It is expected that the roads, parking surfaces, structures, and water would have negative (displayed in red) NDVI values, indicating a lack of vegetation.  As expected, we do see red, negative NDVI values over road and parking surfaces, and structures.  However the surface of LP47 has a range of positive values.  While typical NDVI use dictates a 5-class gradient, 7 classes were used in this case to help identify a more detailed pattern of surface plant health on the pond.  It is evident in Figure 2 that there is healthy algae in the west/southwest corner of the lake.  The circular pattern of healthy algae suggests a source to the south that is potentially entering the lake with some force.

NDVI Algae
 
NDVI Algae Scale

Figure 2: LP47 NDVI Values

 

In addition to processing NDVI data, Pix4D has the ability to generate GeoTiffs, DTMs, and contours from drone data.  Figure 3 was created by layering DTM contours on top of the NDVI values, with the RGB GeoTiff as the bottom layer.  In this case, contours were generated by measuring the elevation at ½ meter X/Y intervals, with ½ meter elevation resolution.  

The layered Figure 3 shows a neighborhood to the south and southwest, at a significantly higher elevation than the lake.  The contours indicate that surface water from the residential neighborhood will likely flow from the neighborhood into the lake, and the gradient of the contours indicates that it is likely the water will flow with some force.

47GeoNDVI.jpg

Figure 3:  LP47 RGB, NDVI, Contour Layers

 

After examining the aerial RGB, NDVI, and DTM contour data for LP47, it would be reasonable suspect nutrient runoff from the residential neighborhood.  Water management technicians may opt to test the runoff entry point in the pond for elevated nutrients relating to lawn fertilizer.  Appropriate mitigation steps may include neighborhood citizen education, discussions with local lawn fertilization companies, and shoreline barriers designed to collect nutrients prior to reaching lake water.

 

Conclusion:

Collection and analysis of aerial RGB and NDVI data can provide efficient analysis of surface water health, including algae monitoring, mitigation, and education.  Aerial data collection is a cost effective, fast, and efficient use of resources to help aid our communities in management of water resources and should be one of many tools available to our environmental resource technicians.  

 

Acknowledgements:

Sentera MultiSpectral Sensor and NDVI Technical Expertise, www.sentera.com.  PAAP Drones, www.paapdrones.com.

 

Contact:

Katie Gilmore, PAAP Drones. katie@paapdrones.com. 612-470-4333.

Comment