Generation of 4 Band CIR Images for Smart Agriculture and Forestry

Generation of 4 Band CIR Images for Smart Agriculture and Forestry

Phase One A/S, a market leader in research, development and manufacturing of integrated medium format camera systems, and JOANNEUM RESEARCH Forschungsgesellschaft MBH, a leading international research organization that develops solutions and technologies for businesses and industry, generated a solution to collect, merge and process RGB and NIR images to enable smart management of agriculture farms and large forests. The solution allows farmers and forest managers to analyze crops and trees health, save water and fertilizer, reduce risk of fire and in general to enhance their work.

About 4 Band CIR

Spongy layers on leaf backsides reflect a lot of light in the near infrared (NIR) domain, in contrast to most non-plant object. The spongy layer collapses in dehydrated or stressed plant and the leaves reflect less NIR light, but the same amount in the visible (RGB) range. Mathematically combining NIR and RGB signals can help in differentiating plants from non-plant objects and healthy plants from sickly plants.

4 Band CIR images are combined red, green and blue (RGB) images with near infrared (NIR) images. Such images are used by farmers and environmental agencies in the assessment of crops and vegetation health.

Unhealthy trees and crops are easily detectable in a CIR image while in a regular image they may still look green.

The Challenge

RGB and NIR images are generated by two cameras that are installed on a rig in the aircraft, taking thousands of images during each flight, in synchronous mode. The cameras may not be of the same type and pixel count may be different between the cameras. The line of sight of the two cameras may differ slightly due to mechanical misalignment. Each lens introduces distortions, which are present in the images and are different between the two cameras. Combining the RGB and NIR images such that each pixel in each of the bands will be correctly placed to form a sharp image presents a huge challenge.

The Solution

Phase One generated the cameras and computer hardware to collect synchronized NIR and RGB images. Test flights were done in various countries, thereby collecting thousands of images.

JOANNEUM RESEARCH enhanced its algorithms and developed new ways to robustly co-register the synchronized images such that the images will keep their original sharpness while adding a NIR band to the Red, Green and Blue bands. This core processing module was used by Phase One to develop mass production software application that enables generation of thousands of CIR images in relatively short time after the collection mission is flown.

Impact and Effects

The 4 Band CIR hardware / software solution is used over the world by farmers for crop health analyses. Vineyards and other high value crops are regularly monitored to enable smart watering and fertilization.

Forest keepers can now easily detect dead trees that hide in the forest and actively remove them. This reduces the amount of potential dry burning material in case of file. Environment agencies and municipal councils can easily estimate the ratio between green areas and built areas.

Multiple other uses in the control and monitoring of plants and green areas are readily available now with an affordable solution made by JOANNEUM RESEARCH and Phase One A/S.

Fig. 1: Combination of RGB and NIR images into a 4 Band CIR Image