- This article is filed under:
- » View All
Food-Agricultural Applications Use Full Spectrum of Machine Vision Cameras
by Winn Hardin, Contributing Editor - AIA Posted 07/29/2015
From saving California wineries from the ravages of extended droughts to finding new ways to feed the world’s growing populations through plant research and precision farming, machine vision is helping companies around the world produce more food with fewer resources.
Starts with a Seed
Allied Vision (Stadtroda, Germany) showed the fruits of its labor with LemnaTec GmbH (Aachen, Germany) last fall in Stuttgart. Allied Vision has provided LemnaTec a variety of visible and shortwave infrared (SWIR) cameras for use in its Scanalyzer digital phenotyping machines and LemnaTec’s larger efforts to help seed producers automate greenhouse plant research on the way to more robust, productive crops. Watch the video here.
“LemnaTec’s application is one of my favorites because they use the full range of our cameras, including visible, SWIR, and long-wave thermal infrared,” says Jean-Philippe Roman, Allied Vision’s manager of corporate marketing. “Customers from multinational agricultural companies to government agencies are using Scanalyzer to research plant growth in large, fully automated greenhouses. The plants are moved around on conveyors to precisely control how much light and water they receive, among other things. And cameras are placed throughout the greenhouse at key stations to measure leaf size, growth rates, and in particular, using our SWIR Goldeye camera, water absorption throughout the plant. Since water absorbs SWIR radiation, the Scanalyzer can take an image of a growing plant and quantify the amount of water in leaves, roots, everywhere…and researchers can measure the effect of droughts and other conditions on plant growth in tightly controlled conditions. The system then displays water concentrations to the researcher using false color images of the plant.”
Hyperspectral imaging systems are helping researchers and food producers measure plant health and growth rates in the field, a practice that may accelerate significantly in Europe starting next year.
“Right now in Europe, we see more call for infrared cameras for hyperspectral systems than for optimized multispectral inspection tools because customers need the full spectral signature to make a judgment on crop health rather than just a few spectral bands,” explains Raf Vandersmissen of Xenics (Leuven, Belgium), a provider of infrared sensors and cameras.
In response to hyperspectral imaging manufacturers’ requests, Xenics recently extended its Xeva line of SWIR cameras to include new type II superlattice (T2SL) detector arrays that extend the camera’s sensitivity out to 2.35 µ. This first T2SL SWIR camera is an area array with 320 x 256 resolution. Additionally, Xenics has recently introduced the Lynx line of SWIR line-scan cameras with up to 2048-pixel resolution for fruit and vegetable sorting applications.
Max Larin, CEO of industrial and scientific camera manufacturer XIMEA (Marianka, Slovakia), says he expects partner companies to start commercially flying Ximea’s HSI compact hyperspectral camera based on IMEC’s new hyperspectral sensor on unmanned aerial vehicle (UAV) drones starting in 2016. “These drone-based services already exist for precision farming, but there hasn’t been a hyperspectral camera small enough or cost-effective enough to fly on a drone for precision farming applications until now,” Larin says. “Our customers are in the process of using our cameras to gather reference spectra now, and we expect commercial applications will ramp up starting next year.”
Commercial interests already are using long-wave infrared thermal cameras on remotely operated drones to measure the water content in grapes growing in vineyards in water-stressed areas, adds Xenics’ Myriam Gillisjans, marketing manager for the Americas. “We have a Canadian customer that’s flying our Gobi-640 long-wave thermal cameras on UAVs to see exactly how much water is needed to keep the vineyard healthy,” says Gillisjans. “With California’s new water restrictions, this has become very important to vineyard operators. We’re seeing similar application in Australia where thermal cameras on UAVs are helping farmers to maintain healthy crops.”
Only the Best Will Do
Measuring water content in the infrared can be useful for more than judging the health of growing plants. It can also help food producers separate good food from bad.
“Fruit and vegetable products basically look for shape, color, and foreign objects,” says Xenics’ Vandersmissen. “When it comes to shape and color, visible cameras work great. But for foreign object detection in frozen peas and corn, for example, SWIR with a water bandpass filter around 1450 nm is great because the water in frozen fruit and vegetables absorbs infrared, while foreign objects appear bright white because they reflect the infrared.”
Measuring shapes, rather than water content, was the goal for Allied Vision’s Prosilica GigE Vision cameras during a recent date inspection application. A date producer used four GC1290C cameras to visibly image individual dates passing along a chute. The images were then analyzed for surface textures called “blisters” that directly relate to date quality and taste. Watch the video here.
Plants need three things to grow: sunlight, nutrients, and water. While sunlight is in abundance, and the earth has a plentiful supply of nutrients, some developed parts of the world are starting to experience water shortages traditionally limited to emerging markets and arid climates. In response, countries around the world are looking for ways to use agricultural resources more efficiently. And efficiency requires research and measurement, both of which can benefit from automation. Machine vision technology, in the lab and the field, will help humanity ensure a fruitful and healthy future.
There are currently no comments for this article.
Leave a Comment:
All fields are required, but only your name and comment will be visible (email addresses are kept confidential). Comments are moderated and will not appear immediately. Please no link dropping, no keywords or domains as names; do not spam, and please do not advertise.