Search for Optimum Color Space for the Recognition of Oranges in Agricultural Fields

Artificial vision systems are powerful tools for automatic recognition of fruits and vegetables in Mexican agricultural fields. This research presents the image capture, cropping and process for the recognition of fruits found in trees. The design and implementation of a recognition system based on color is proposed. For this research, experiments were carried out with the following color spaces: Hue Lightnes Saturation (HLS), Hue Saturation Value (HSV), Luma Chrominance-red Chrominance-blue (YCrCb), Luminance Chrominance-U Chrominance-V (YUV), Luminance red/green yellow/blue (L * a * b *), Luminance Chromaticity-u Chromaticity-v (L * u * v *), Tonality Saturation Lightness (TSL), Intensity-red Intensity-green Intensity-blue (I1I2I3) and XYZ. In each color space, different alternatives emerge, for example: what channels to use, reduction of the image, size of the histograms, etc. A varied set of images were selected to test these techniques for the color recognition of orange fruit. The results showed that some color spaces are the most appropriate for the recognition of oranges.