Chemists are turning their smartphone cameras into real-time chemical sensors, replacing the invasive probes that currently slow down manufacturing. Most chemical manufacturing relies on probes that must be physically inserted into a reaction vessel, and each probe can only monitor one process at a time. This creates bottlenecks: a chemist developing a new drug might have to run dozens of reactions sequentially because they lack enough probes to watch them all simultaneously. Meanwhile, many chemical reactions produce visible colour changes that trained eyes can spot but no instrument routinely measures. The problem is that existing monitoring technologies work well at either lab-bench scale (a teacup) or industrial-plant scale (a swimming pool), but not both. This project aims to build computer-vision software that lets any standard camera—from a lab webcam to a phone—quantify colour changes in chemical reactions without touching the process. If successful, it would let chemists monitor dozens of reactions at once, speeding up product development and reducing the need for expensive, single-purpose hardware. The global market for digital process monitoring in chemicals is expected to reach $31 billion by 2028, and this work targets the fundamental software layer that would make that shift practical.
View original technical description
KEYWORDS: chemistry, computer vision, image processing, manufacturing, productivity, process monitoring. Chemical and biochemical manufacturing are dominated by colour changes, both subtle and stark. Such phenomena are often reported by-eye but not routinely quantified, especially over time. This renewal of a research and leadership programme aims to empower any chemist with any camera to capture any visible trend from any high-value chemical process, all without having to disturb the process under study. Most industrial chemists are accustomed to extracting chemical monitoring information using invasive, probe-based technologies. These technologies are robust and trusted. However, no current technologies are seamlessly applicable to monitoring chemical processes in real-time on BOTH the high throughput lab scale (the 'teacup') AND process/plant scales (the 'swimming pool'). Instead, current process analytical technologies are oftentimes tied to one specific hardware platform, and each example of such probe-based hardware can only monitor one process at a time. Ultimately, this can produce bottlenecks in analysis, slowing chemical product development and deployment. To address this productivity and chemical data throughput challenge, there is a real drive from R&D budget holders to invest in digital-ready analytical technologies. Computer Vision is the science of digitally quantifying real-world colours and objects using cameras. With cameras and computer vision, and further development through this fellowship renewal, the hardware and software needed for more time-, cost-, and safety-effective monitoring of high-value chemical processes can be realised in an accessible and globally adoptable manner. The global investment for digitalisation of process analytical technology (PAT) in the chemical industry is expected to reach $31 billion by 2028, representing an annual growth rate of approximately 6% from the present $23.5 billion market (sources: Made Smarter Review, 2017. Frost & Sullivan, 2017, and 2022). Underpinning this trend, R&D managers across chemical manufacturing are driving the streamlined adoption of new digital-ready chemical technology, to improve productivity, process safety, and ability to exploit the adjacent evolution of artificial intelligence.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know