As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.
An inspection camera works by carrying its own light down a cable, bouncing it off a surface a few centimetres away, and converting the reflected light into a video signal on a small image sensor sitting behind a fixed focus lens. The cable sends power out to the diodes and the sensor and brings the video back; the handheld unit decodes it onto a screen or a wireless module streams it to a phone. Nothing in that chain measures distance, size or temperature, which is why a thermal imager sold under the same name works on a completely different principle.
From LED light to pixels: the signal path
Follow one photon and the whole device makes sense. A diode at the tip emits visible light. It travels a few centimetres, strikes the inside of a pipe or a cylinder wall and scatters. A fraction of that scattered light re enters the lens, which projects it onto a silicon sensor divided into a grid of photosites. Each photosite accumulates charge in proportion to the light it received, the charge is read out and digitised, and the resulting frames are serialised and pushed back along the cable. At the far end a processor turns the stream into a picture and paints it on a panel.
The numbers quoted in listings attach to different links in that chain, which is why they look contradictory. The NIDAGE wireless endoscope states a 2.0 megapixel lens with three selectable resolutions of 1920 by 1080, 1280 by 720 and 640 by 480, which is a statement about the output format rather than the optics. The DEPSTECH DS300 states 2.0 megapixel stills and 1080P video on a 4.3 inch panel giving a horizontal viewing angle of about 170 degrees, and separately states a second generation sensor chip adopted to raise the recording frame rate and remove picture delay. Frame rate and delay are processing properties, not optical ones, and improving them does nothing for sharpness. A native sensor pixel count is not stated in either listing, so the output format is all you have to compare. Sensor, output format and panel are three independent figures, and the weakest one sets what you see.
Fixed focus optics and why the working distance is so short
Depth of field grows as aperture shrinks and as focal length shortens, and a probe lens is extreme on both counts: an aperture under a millimetre and a focal length of a couple of millimetres. That combination gives a usable band of a few centimetres around the design distance, which is enough to be useful and far too little to be forgiving. The NIDAGE wireless listing states a focal length of 3 to 10 cm and notes that the instrument is mainly for viewing close objects, which is an honest description of the constraint rather than a limitation of that particular unit.
Two practical consequences follow. First, pressing the probe against the fault guarantees a blur, because the surface is now inside the near limit. The correct action is to withdraw until the image snaps in, which on most of these is a retreat of two or three centimetres. Second, large voids are poor subjects. In a wall cavity the opposite leaf may be 100 mm away, well beyond focus, so you get a sharp view of the near face and a soft wash behind it. Knowing that in advance stops you blaming the tool for physics.
Why the cable is both a conduit and a positioning device
Electrically the cable is unremarkable: a few conductors for diode power, sensor supply and the video stream, inside a sealed sheath. Mechanically it is the most important part of the tool, because a fixed lens can only be aimed by moving the whole probe. This is why these cables are semi rigid rather than floppy. You pre bend a hook into the last few centimetres, feed the probe along the route, and rotate the cable at your hand to sweep the view.
The engineering compromise lives at the joint between the head and the cable, which is where sealing, flexing and conductor strain all meet. It is the usual failure point, and the symptoms are electrical rather than optical: noise bars, colour shifts or dropouts that appear exactly when the cable is flexed at that point. If the picture degrades on movement and recovers when still, suspect the joint rather than the sensor. Length matters too, since a longer run means more voltage drop to the diodes at the tip, so a 16 ft cable with the brightness at maximum does not deliver the same illumination as a 5 ft one with the same diode count.
Wired screen against a wireless phone link
Two architectures exist and they feel completely different in the hand.
| Behaviour | Wired handheld unit | Wireless unit with a phone |
|---|---|---|
| Signal path | Cable to a local processor and panel | Compression, local wireless link, app decode |
| Latency | Effectively immediate | Variable, enough to make steering awkward |
| Setup | Power on and work, with no pairing | Join the access point, grant permissions, open the app |
| Screen size | Fixed, typically 4.3 inch | Whatever the phone offers |
| Recording | Only if the unit states it | Handled by the app, so usually yes |
| Longevity | Independent of software support | Depends on the app surviving operating system updates |
| Hygiene | Keeps the phone out of the engine bay | Phone goes where the work is |
The hygiene row is not a trivial point and the manufacturers raise it themselves: the DEPSTECH bullets promote the onboard screen on the grounds that you need not dirty an expensive phone at work. Against that, a phone screen is larger, brighter and already in your pocket. Our guide to choosing a borescope for inspection work treats this as a workflow decision rather than a quality one, and the NIDAGE inspection camera review covers how the wired and wireless versions of the same probe differ in use.
A different sensor entirely: how a thermal imager sees
A thermal camera shares the word camera and almost nothing else. There is no illumination, because the subject supplies the signal: every surface above absolute zero radiates long wave infrared, and the amount depends on its temperature and its emissivity. The lens is germanium rather than glass, since glass is opaque at those wavelengths. Behind it sits a microbolometer array, a grid of tiny thermally isolated elements whose electrical resistance changes as they absorb that radiation. The electronics read the resistance map, apply the emissivity and distance settings, and produce a temperature for every pixel.
The Flagfront YXI96 thermal camera states the figures that matter in that chain: 240 by 240 super resolution imaging, a 25 Hz refresh rate, a measurement range of -4 to 1022 degrees Fahrenheit with an accuracy error within 2 percent, and adjustable distance and emissivity. Those last two adjustments exist because the sensor is inferring temperature from radiated energy, so it needs to know how efficiently the surface radiates and how much atmosphere sits in the way. A borescope needs no such settings because it is not inferring anything. The practical division is simple: a thermal imager finds the hidden anomaly through the surface, and a borescope then goes in through the smallest hole to see what it is. Our guides to choosing a thermal camera for home inspection and what NETD means on a thermal camera cover the sensitivity figures behind that first step.
Using the mechanism to read a blurry image
Every common complaint maps onto one stage of the chain, which makes field diagnosis quick.
- Soft at close range. The target is inside the near focal limit. Withdraw a few centimetres rather than pushing in.
- White glare in the middle. Diode light is reflecting straight back from a shiny or wet surface. Reduce brightness and approach at an angle.
- Dark beyond the near field. The diodes have run out of reach. There is no fix other than getting closer, because illumination falls away quickly with distance.
- Grainy and dim everywhere. The sensor is amplifying a weak signal. Expect this from the slimmest probes, which have the smallest lens and the fewest diodes.
- Stutter or lag while steering. A processing and link problem, not an optical one, and the reason a wired screen is easier to aim with.
- Bars, colour shifts or dropouts when the cable moves. The joint at the head is failing. Nothing on the screen will fix it.
Used this way the tool becomes predictable. You know it cannot focus, so you control distance. You know the light comes from beside the lens, so you control the angle. You know the cable aims the lens, so you pre form the bend before entering the hole. None of that is in the instructions, and all of it follows from the signal path.
Frequently Asked Questions
Does the camera have a light of its own?
It must, because the spaces these tools enter have none. A ring of light emitting diodes surrounds the lens at the tip, and the brightness is adjustable from the handle. That geometry is also the source of the washed out picture people see at very close range, since the light and the lens are pointing at the same spot from almost the same place.
Why can the lens not be focused?
There is no room for a focus mechanism in a probe a few millimetres across, and no sensible way to drive one from the far end of a cable. The lens is cemented at a fixed distance, giving a short usable band in the region of 2 to 10 cm on these products. Focusing is done by moving the probe.
What limits the resolution on the screen?
Three separate numbers, which listings often blend. The image sensor has a native pixel count, the processing chain outputs a video format, and the display has its own panel resolution. A unit may state a 2.0 megapixel sensor, a 1080P output and a 4.3 inch panel, and the smallest of the three is what you actually see.
Why is the wireless version laggy?
Because the video is compressed at the probe end, sent over a local wireless link and decompressed by an app before it reaches the screen. Every step adds delay, and the delay is variable, which is worse than a fixed one when you are steering a probe by hand. A wired handheld decodes locally and feels immediate.
How does a thermal camera differ inside?
It has no illumination at all. A microbolometer array absorbs long wave infrared radiated by the scene, each pixel changing resistance as it warms, and the electronics convert that into a temperature map. It sees a surface pattern from the outside rather than an image from inside a cavity, and it needs an emissivity setting because different materials radiate differently at the same temperature.
Does the cable affect image quality?
It can. The video is a low voltage signal travelling several metres beside the power leads for the diodes, so a damaged or badly joined cable produces noise, rolling bars or intermittent dropouts. Those symptoms appear when the cable is flexed, which is the quickest way to tell a cable fault from a sensor fault.