Deep Empirical Observations

April 26, 2021

How can we know better than our competitors? The quality of our empirical observations determines the quality of our work, whether searching for secrets or monitoring for risk factors.

Digitisation is the art of transforming the atoms of the physical world into the bits (or pixels) of the digital realm. This is a necessary prerequisite to many analytical and automated workflows. Think of gold exploration, where we rely on the sensors to elucidate the mineral of interest from the sky, often beyond the capabilities of human vision. If the quality of our empirical observations determines the quality of our work, whether searching for secrets or monitoring for risk factors, then it is the quality (and sensitivity) of our sensors that frames the most important question: how can we know better than our competitors?

In the domain of empirical analysis the quality of data defines the quality of work. Scientists may have theorised many important questions long before they were answers, but it is only when sensors convert atoms into bits, or pixels. As Richard Feynman once said: "Everything is made of atoms." In many lines of scientific and empirical inquiry, everything is being replicated in the digital realm with impressive resolution. Among the most remarkable applications is the ability to spot diffuse methane gas leaks from space. Hard to imagine a more stunning scientific achievement.

The human vision system is an impressive bit of engineering. Well-trained agronomists can spot pests in the field. Viticulturists spot leaf and stem disease. Physicians are masters of screening for subtle changes in a patient. In some cases, direct observation does the trick. However, what if we want to sort coffee or cocoa beans not by supplier or country of origin but by the flavour profile of individual beans? What if we want to map brain cancer during surgery in real-time? What if we want to predict the freshness level or sugar content of fruit and vegetables, at the distribution centre or at the retail location? Visual inspection and the smartphone can take us only so far.

If the smartphone is, as Elon Musk says, turning humans into "cyborgs" then what of hyperspectral imaging? Hyperspectral cameras combined with advanced artificial intelligence can rapidly and non-invasively scan individual units across a range of parameters (e.g. sugar, cancer, protein) for zero marginal cost. For organisations, operators, scientists, engineers and clinicians who are paid to know their objects of analysis, this is a powerful tool to transform hypotheses and risk into high confidence predictions so that you can unlock maximum value.

Book a consultation to learn more

We can walk you through how hyperspectral imaging can improve your workflows.