April 23, 2019

How To Boost Hemp Production Using Multispectral Imagery

In this episode of American Cannabis, MicaSense and Arbor Drone discuss how hemp farmers can gain valuable insight using multispectral imagery with the help of drones.



Ellis Smith, the host of American Cannabis, welcomed Drew Baustian, Business Development Manager from MicaSense, along with Dan Staley, Principal at Arbor Drone, for an in-depth look at the value in aerial imagery for hemp farmers and the potential of multispectral images for detailed crop analysis. With a degree in Agricultural Science from Cornell University, Baustian has been working with MicaSense for over three years and brings previous experience from the agriculture industry. Staley, a skilled UAV pilot, and well-published researcher studied Environmental Horticulture and Urban Forestry at the University of California, as well as, Urban Planning at the University of Washington.

The Rise of Technology in Precision Farming

Unmanned aerial vehicles (UAVs), more commonly known as drones, are finding their place in precision agriculture. Staley suggested we are, “seeing more drones in precision ag due to technological breakthroughs – drones are cheap, fast, and very robust.” In fact, the use of UAVs in agriculture is expected to top $1 billion globally by 2024. However, Staley also reminded, UAV pilots must be licensed to fly their drones for agriculture or commercial uses, and it is critical to know and understand the restrictions enforced by the FAA.

Integrated Sensors

Technology assistance comes in many forms in agriculture including guiding equipment, planting, spraying, and harvesting. Various smart sensors are being used all over the globe to monitor plant health for improved yields and reduced costs. Agronomists interpret the data collected by sensors such as:

  • Soil moisture, temperature, humidity
  • Spectral sensors flown by aircraft, drone, or satellite

Together, these sensors and images help cultivators increase yield and yield quality. Plus, farmers benefit from identifying specific areas in the field which need the most attention, catching pests early, and scouting for weeds.

MicaSense and Drone-Based Multispectral Imagery

Operating in more than 70 countries, MicaSense is the standard in drone-based multispectral imagery, offering a suite of sensors for agriculture uses including their RedEdge-MX camera. The company currently supports both academic and commercial research, government agencies, and the growing demographic of high-value crop farmers and managers.

Explaining how the sensor works, Baustian said, “[the sensor] captures the reflectance of five spectral bands: blue, green, red, red edge, and near infrared.” He continued, “the raw images are black and white, showing brightness or darkness as it relates to the reflectance of the plants.”

 

Explaining Reflectance

Plants reflect and absorb light, and it is visible throughout their growing cycle. During the growing season, plants absorb red light and reflect green light. In the fall, when the leaves change, this is due to a decrease in red absorbance and more reflectance. The MicaSense RedEdge MX highlights this reflectance in crops. A dark image indicates a healthy, green plant, while lighter areas may indicate the plant is stressed or even dying.

“Using a plethora of data sets, with the different images, an agronomist gets an understanding of what’s happening in that canopy,” Baustian stated.

In the vineyards, MicaSense imagery has helped identify failed shoots on grape vines and allow farmers to target specific areas rather than scout an entire field. Similarly, Baustian showed how sensors might be able to help cannabis cultivators identify the difference between male and female plants and reduce the labor costs of manual scouting. Although not yet perfected for plant sexing, UAVs and MicaSense imagery can help with,

  • Plant inventory and harvest forecasting
  • Plant size and harvest potential
  • Sparse plants or troubled areas in large fields

Machine Learning

By teaching computer programs how to learn and what to look for in making an analysis, Staley explains how machine learning is a growing technology easing the grower’s work. “As machine learning catches on and computing power increases, we can expect to see machine learning on the rise,” Staley said. “We’ve taken vegetation indices and taught the computer what a healthy plant looks like and what an unhealthy plant looks like.” Staley went on to show how the software can help identify and declutter the images by removing weeds from the background for accurate analysis of the fields.

Making the Final Decision

In summary, Staley reminds that ROI is the most crucial consideration. Farmers need to determine whether the cost per acre for this type of analysis is worth the return. Additionally, cultivators need to consider whether to buy the equipment and meet all the requirements for licensure or hire a trained pilot already licensed. “With the returns possible with hemp, it is probably a good investment to wrap your mind around your field,” Staley suggested.

To learn more about MicaSense sensors such as RedEdge-MX or the new Altum, or the integration kits necessary for drone mounting, visit their website at www.micasense.com.