In this issue of Plant and Soil Nakaji et al. (Plant Soil, this volume, 2008) report a novel approach for automatically identifying roots and other rhizosphere components in rhizosphere images acquired using a multi-spectral (visible—VIS- and nearinfrared—NIR-) imaging system. The images are acquired through a root-window observation device and the study highlights the perspectives offered by this imaging system. An outstanding outcome of this research is that the new approach can be applied to effectively separate soil litter from the purely mineral phase and distinguish root tissues that differ in physiological status, i.e. live (different age classes), senescent and dead. If achievable routinely, such a detailed classification of rhizosphere components could greatly improve our appraisal of root turnover and associated organic matter input to the soil, information of paramount importance for an improved understanding of many essential processes such as global geochemical cycles. Minirhizotrons (MR) systems have been increasingly used in global change studies because they are a convenient way to frequently and nondestructively quantify root length production and mortality (Norby and Jackson, New Phytol, 147:3–12, 2000; Hendrick and Pregitzer, Ecology, 73:1094–1104, 1992). However, the MR technique still has many limitations, including the lack of a standard, accurate and rapid procedure to extract and classify rhizosphere components from the MR images obtained. The recent work by Nakaji et al. (Plant Soil, this volume, 2008) provides convincing evidence that the inclusion of a VIS-NIR multispectral capability into conventional MR systems could substantially improve this method, and extend its adoption by the wider plant scientist community as a standard research tool.