Thermal vs image intensification: choosing the right sensor for the mission

The framing of thermal versus image intensification as a choice is already somewhat dated, but it remains the right place to start because it exposes the assumptions that drive most sensor procurement. The two technologies do not compete on the same axis. Thermal imaging, whether cooled or uncooled, builds a picture from emitted infrared radiation. It sees heat, which means it works in complete darkness, through smoke, dust, light fog and many camouflage screens, and it does not care whether the target is lit. Image intensification does the opposite. It collects the small number of photons available at night, in starlight or urban glow, and multiplies them. That produces a scene with far more texture, edges and facial detail, but it collapses in total darkness and is degraded or defeated by bright light sources, smoke and obscurants.

The practical consequence is that the question is not which sensor is better but which task the user is trying to complete. Detection, recognition and identification are different problems with different optical and spectral requirements. Thermal tends to win at detection, especially at range and through obscuration. Image intensification tends to win at recognition and identification at shorter range, where the human visual system needs the spatial detail that a thermal core often smooths away. A thermal image of a person at distance is a hot blob with a shape; an intensified image of the same person at closer range can show clothing, posture and a face. That distinction drives doctrine, rules of engagement and the legal review that sits behind them.

The mechanics also explain the cost and logistics split. Cooled thermal cores offer better sensitivity and range but need cryogenic cooling, which adds weight, power draw, cooldown time and maintenance burden. Uncooled microbolometers have closed much of the gap for many tactical ranges and have become the default for small UAS, weapon sights and helmet-mounted systems because they are lighter, cheaper and instant-on. Image intensifier tubes are a mature, comparatively affordable technology, but they are also a constrained supply chain. High-performance Gen III and filmless tubes depend on a small number of photocathode and microchannel plate producers, and demand has historically outrun capacity. That supply reality, more than any performance argument, is why many programs now treat intensification as a scarce resource to be allocated rather than a default fit.

The demand side is being reshaped by two forces. First, the small UAS and loitering munition market has standardized on thermal and low-light CMOS because those sensors are compact, rugged and cheap enough to be attritable. A drone that costs a few thousand dollars cannot carry a cooled mid-wave infrared core, and it does not need to. Second, the dismounted soldier and vehicle crew market has moved toward fusion. Fused thermal and intensified imagery, or thermal plus low-light color CMOS, gives the operator both the detection advantage of heat sensing and the situational detail of reflected-light imaging. That fusion is now a software and display problem as much as a sensor problem, which is why the value is migrating toward integrators, image-processing pipeline vendors and display suppliers rather than remaining with the tube and core makers alone.