Detailed Imaging and Pixel-Level Accuracy with Emerging Image Sensors
Emerging image sensors are being developed with capabilities beyond well-established CMOS detectors, with the possibility of detecting broader parts of the light spectrum that are beyond human vision. IDTechEx’s report, “Emerging Image Sensor Technologies 2024-2034: Applications and Markets”, covers new technologies including hyperspectral imaging and event-based vision. New image sensing technologies are able to gather spectral data on a pixel level, which can increase both resolution and dynamic range, as they carry out imaging over a larger area.
Hyperspectral imaging and agriculture
Hyperspectral imaging is one of the main emerging image sensor technologies covered in IDTechEx’s report. Each location in an image is represented by a spectrum, which can achieve a more detailed view of an object’s composition (a 3D data set known as a hyperspectral cube) when compared to traditional red, green, and blue pixels. Point scan, line scan, wavelength scan, and snapshot are all device architectures used for hyperspectral imaging, and each uses varying approaches and light distributions to achieve different images. These approaches are explored in detail in IDTechEx’s report.
Agriculture is the main application for hyperspectral imaging, making up for more than half of the uses for this technology. Detailed information on the health and status of plants can be obtained with hyperspectral imaging to identify any issues with crops that may predict the harvest quality. With the never-ending demand for crops such as wheat, hyperspectral imaging can be extremely important for the economic security of farming as an industry and can help with precision agriculture.
A graph showing the importance of hyperspectral imaging qualities by sector. Source: IDTechEx
Event-based sensing and dynamic vision sensing
Event-based sensors are able to pick up whenever the signal changes with the use of a pre-defined threshold and include a timestamp. Some advantages of general event-based sensing include lower power consumption, increased dynamic range, reduced computer processing requirements, and reduced data transmission.
Event-based vision or dynamic vision sensing (DVS) reports timestamps alongside percentage intensity changes at the pixel level to provide greater resolution and range. It is described as a subset of event-based sensing in IDTechEx’s report and is a method that mimics how the human brain sources information, recording changes as they occur instead of recording an entire frame at intervals. DVS can produce greater temporal resolution and less data, meaning processing can be much simpler. However, no color information can be detected, and datasets can be difficult to interpret without specialist software.
Machine vision, or computer-based image analysis, is the primary driver for emerging image sensors such as event-based sensors. The more data that machine learning can acquire, the more accurately objects can be identified and classified. IDTechEx reports that optical data from a different wavelength range or with pixel-level resolution will be beneficial to a number of growing technology sectors.
IDTechEx’s closer look at emerging sensing technologies
Image sensing based on quantum dots detecting wavelengths beyond what is possible with current exploration techniques, miniaturized spectroscopy, wavefront imaging, x-ray image sensors, and computer vision, are all additional emerging image sensing technologies detailed at great length within IDTechEx’s latest report, “Emerging Image Sensor Technologies 2024-2034: Applications and Markets”. Downloadable sample pages are available for this report.
For the full portfolio of sensors, haptics and displays market research available from IDTechEx, please see www.IDTechEx.com/Research/Sensors.
