Creating a panoramic field image using multi-spectral stereovision system

Satellite-based remote sensing presents a broad view of field scenes while ground-based remote sensing offers detailed observation of the crop to support precision farming decision-making. If it is possible to combine the advantages of both satellite-based and ground-based sensing, the resulting imagery could present both the broad view of a field scene and detailed observation of crop growth, which offers a very informative means for delivering crop production information. This paper describes a fundamental investigation on creating multi-spectral 3D panoramic field imagery by seamlessly integrating ground-based multi-spectral images of the field. A vehicle-mounted stereovision sensor composed of two multi-spectral cameras collects six-spectral channel images to form the multi-spectral 3D field imagery. The collected stereo image stream is then synthesized progressively to create a single seamless field image. The resulting panoramic imagery forms a 3D virtual field scene capable of providing more informative interpretation of crop/field status than a 2D plane image. Field test results proved that the assembled multi-spectral imagery was capable of providing a seamless 3D panoramic view of a soybean field with a broad view for assessing crop growth status with sufficient detail of individual crop plants.