Early apple bruise recognition based on near-infrared imaging and grayscale gradient images

Early apple bruises, especially those occurring within half an hour, usually have no external symptoms and are difficult to recognize. This study developed a fast nondestructive detection method for early bruises based on near-infrared camera imaging image recognition. A total of 31 samples were photographed both sides each apple. Grayscale images the sound apples captured using with wavelength region between 900 2350 nm. Images (n = 62) without collected. The identical artificially damaged at two stages (0 h, 0.5 h) by camera, 186 grayscale As light spot surface limits accuracy in detecting defects, compound pre-processing was proposed consisting nonlinear transformation frequency-domain filtering techniques. Finally, gradient is obtained taking first-order derivative preprocessed image. Since bruise had distinct edges, more favorable identification. 97.62% classification nonbruised fresh bruises. experimental results show that it feasible identify images. In subsequent study, will be further improved, parameters involved algorithm could adjusted adaptive variables. addition, system explored suitable online detection.