Discrimination of Agricultural Crops by Infrared Temperature Pattern Processing

Abstract

Pattern discrimination of agricultural crops using computer on-line operated infrared thermal camera is reported. The camera has the capability of temperature resolution of 0.05deg by the liquid nitrogen cooled mercury cadmium telluride sensor. By the computerized pattern acquisition system, infrared thermal signal is directly digitized and transferred to the core memory of the computer and then restored in the magnetic tape file. The full graphic digital color display system of raster scan format with the picture element of 256×256 or 512×512 displays the processed thermal pattern by real, pseude or hybrid color display format.
Favorable extraction of the orange fruits from the branches and the leaves is obtained by the multi-level thresholding of their temperature.
Discrimination of the vegetable (Komatsuna) and the soil surface is also made by the same method.
This infrared pattern recognition technique is expected to be a powerful tool for the visual aid to automated farming system and the instrumentation or selection of agricultural crops.