USE OF IMAGE FOR CLASSIFICATION AND IDENTIFICATION OF FOREST SEEDS QUALITY

Abstract

Seeds are essential inputs in agriculture and forestry, being the viability, crucial criteria for determining their quality. Image processing and analysis methods were explored to assess seed quality. The vigor of seeds is vital for the establishment of crops, being defi ned as their ability to germinate and grow in diff erent environments, often inhospitable. Conventional methods, such as germination tests, tetrazolium and the accelerated aging test, have their limitations to determine aspects of predictions and vigor, promoting the search for non-destructive and automated approaches. Image analysis, including near-infrared spectroscopy (NIR) and hyperspectral imaging (HSI), is an eff ective tool for evaluating various seed characteristics such as size, color and shape. Equipment such as GroundEye® allows for accurate and quick analysis of seed quality. Image analysis has expanded rapidly, off ering alternative options to ensure the determination of seed quality of forest and agricultural species, with the potential to revolutionize the selection of plant varieties and increase productivity.