Comparison of Image Data Obtained with Different Commercial Cameras for Use in Visible Spectroscopy

Abstract. Image processing techniques have been widely used in various applications in different fields including engineering, medical, forensic, military, and agriculture. Color images taken for further analysis have now become predominant focus specifically in agricultural application. The main objective of this study was to compare processed image data of color images of the same targets captured by different commercial cameras under different lighting conditions. The experimental design comprised five cameras (Brands: Canon, AEE, PYLE, Bushnell, and LG) by three targets (a piece of black-and-white checkerboard cloth, an apple fruit, and a bunch of holly bush leaves) by four light settings (300 W, 600 W, 900 W, and ambient light). Each image was cropped and resized for consistency before data from R (red), G (green), and B (blue) bands of the images were extracted/exported to Excel by using the new ‘RGB2X 1.0‘ software application. The extracted data were processed into ‘Adjusted Green Red Index‘ (AGRI) dataset for further analysis. Histograms of the R, G, B, and AGRI datasets were obtained and compared. The results indicate that for the checkerboard images, the Bushnell and the LG cameras produced comparable mean modal AGRI values. For the apple images, the following camera pairs gave comparable mean modal AGRI values under certain lighting conditions: AEE and PYLE; AEE and Canon; and Bushnell and LG. Lastly, for the leaves, the following camera pairs gave comparable mean modal AGRI values: the AEE and PYLE cameras; and the Canon and LG cameras. These results demonstrate the potential of obtaining similar outputs from processed images acquired by different commercial cameras and provide insight into the ability to use certain low cost and handy cameras to obtain similar outputs as the expensive cameras.