Potential of satellite hyperspectral imaging technology in soil health analysis: A step towards environmental sustainability

Due to extensive anthropogenic activities, soil fertility may degrade severely, threatening the food security. Traditional soil sampling from fields, and their subsequent chemical analysis, is invasive, time-consuming, and not economical at the regional or global level. Remote sensing techniques, specifically the hyperspectral imaging due to their sensitive narrow wavelength bands, offer a non-invasive, fast, reliable, and economical technique to map soil health parameters at the regional scale. New generation hyperspectral satellites with better signal to noise ratio (SNR), may open up a new domain of rapid soil health digital mapping to meet the sustainable precision agriculture targets under the UN SDGs. Despite the high potential, because of the non-availability of satellite sensors (beyond EO-1 Hyperion with low SNR), and limited expertise in high dimensional data handling beyond the scientific community, hyperspectral imaging has restricted applications in agriculture. This review summarises the applications of satellite hyperspectral imaging for soil health assessment, and the developed models. It identifies the research gaps for wide-scale soil and agricultural applications using new-generation hyperspectral satellites. It also examines the upper hand of hyperspectral over multispectral images for assessment of soil health, and critically analyses the various satellite hyperspectral sensors for assessment of soil health parameters, as an efficient alternative to traditional field-based methods. Finally, the review identifies the challenges in the large-scale application of the technology and the way forward for popularisation of hyperspectral imaging for ushering in environmental sustainability. This extensive compilation of reports for assessment of soil attributes through satellite hyperspectral imaging would eventually help researchers to focus on the grey areas, with possibilities to integrate cutting-edge AI/ML models with latest hyperspectral satellites’ datasets for a wide range of soil and agricultural applications.