Thermal Master Thor002 Infrared Camera Review – Price breaker for technology freaks and semi-professional thermal imaging at a competitive price

1 - Introduction, Unboxing and Technical Specifications 2 - Functions and Menu Navigation 3 - Practical Use and Evaluation Software 4 - Summary and Conclusion

Measuring principle and sensor technology of the Thor002

The Thor002 is based on an uncooled microbolometer with a native sensor resolution of 256 x 192 pixels. This results in a total of 49 152 measuring points, which are output as a thermal raster image. Each of these points corresponds to a detector element that operates in the long-wave infrared range between 8 and 14 micrometers. In contrast to cooled sensors, a microbolometer uses temperature-dependent resistance changes in an absorbing material (usually vanadium oxide or amorphous silicon) to detect thermal radiation. Since no active cooling is required, costs and size remain low, but at the expense of lower thermal sensitivity and higher noise density.

The Thor002 also features a visible laser that marks the center of the thermal target area. However, the actual temperature measurement is not carried out via this laser, but via the corresponding central sensor element in the bolometer matrix. The laser is only used as an optical reference to ensure that the user is aiming exactly at the point whose temperature they are interested in. The accuracy of the spot measurement depends on several factors, including the ratio of the object size to the measuring spot (distance-to-spot ratio), which in turn results from the optics and sensor resolution. As a single pixel covers a larger area on the target object at greater distances, the measurement result is increasingly falsified for small objects or imprecise distance specifications.

The distance to the target object can be entered manually in the Thor002. This information is incorporated into the software correction of the temperature measurement. The background to this is the fact that heat radiation is attenuated with increasing distance due to air absorption and scattering. This can lead to significant deviations, particularly at greater distances or with high humidity. The distance input allows the internal signal processing to compensate for this effect and thus calculate a more realistic surface temperature. This correction is particularly important for precise analyses of pipelines, facades or electrical installations.

The Thor002 combines the thermal image with an additional visual image from the integrated digital camera. This is either stored separately or, if required, combined with the IR image as an overlay. The visual supplement provides orientation and helps with documentation in particular, as thermal anomalies can be better localized. Without a visual image, it would often be unclear exactly which object was examined, especially as thermal contrasts alone cannot always be clearly interpreted.

Super Resolution in practical comparison and repeatability

The following two IR images illustrate the effect of the super resolution function. According to the manual, the Thor002 uses a process in which several slightly offset individual images are combined to produce a thermal image with a higher effective resolution. Edges are smoothed, noise is reduced and small temperature differences are better displayed. In practice, this means that the standard image appears coarser and shows more noise on homogeneous surfaces, while the super-resolution image shows finer structures, such as pipes, cables or joints, more clearly. The measured temperature value itself does not change, but the visual significance and the ability to recognize smaller details thermally does. This function is particularly useful when analyzing weak points in insulation, water damage or electrical installations where localized overheating can occur. The differences in the pictures are the result of my unsteady hand.

A clear weak point of the Thor002 is the lever in the handle, which is designed as a trigger. Although this so-called trigger seems mechanically stable, it is clearly too stiff to operate. As a result, it is not uncommon for unintentional movements to occur when the shutter is released, which can impair image sharpness when used hands-free. In addition, this lever does not have a clearly defined tactile release point. There is no perceptible resistance or slight mechanical feedback to signal to the user exactly when the shutter is released. Instead, the transition from light pressure to actual activation is blurred, which is particularly annoying when taking precise shots or when using the device on a tripod. For a device that is also used in the semi-professional sector, a clean, smooth-running and clearly latching release mechanism would be desirable. This would not only improve handling, but also increase the repeat accuracy of continuous shooting.

Practical examples and sources of error

The first image shows the drum section of a front-loading dryer. The round shape of the drum, which is clearly visible in the infrared image, stands out from the rest of the housing due to its lower temperature. While the housing has a maximum temperature of 24.1 °C according to the measurement, the center of the drum remains significantly cooler. This indicates a good thermal separation between the interior and the outer casing. Such a difference can be used to assess the insulation of the appliance. Heat losses would be easy to detect here, particularly during operation, which could indicate defective insulation or blocked air ducting, for example. The cardinal error here is the oblique angle, which makes the values appear too low.

The second picture shows a typical electronic household electricity meter in the switch cabinet, this time at the correct angle. The maximum temperature here is 19.2 °C, with the heat development concentrated in the upper connection area. The central area of the meter housing remains at a largely uniform temperature. This indicates largely low-loss signal and energy processing inside the meter itself. However, the localized heating at the upper cable connections is noticeable. This could be due to increased current consumption in combination with tight installation space or inadequate ventilation. Such a measurement is particularly recommended for older installations or if overheated terminals are suspected and can provide indications of thermal weak points in the power grid as part of preventive maintenance.

Both examples illustrate the practical benefits of a thermal imaging camera such as the Thor002 for quick and non-contact condition analysis in the home. They can also be carried out without interfering with the appliance or the installation and provide usable results in real time for assessing operational safety and energy efficiency.

Limits of emissivity measurement directly in the camera

A crucial aspect of thermographic analysis is the correct assessment of surface emissivity. Particularly with metallic or shiny surfaces, the temperature values can be significantly misinterpreted if you rely on the automatically generated color scales of the camera. In the image shown here of an open door, the door leaf appears warmer than its surroundings, whereby reflective components such as the door handle or possibly bare metal frames are incorrectly interpreted as colder. The reason for this is the low emissivity of such materials, which emit less heat radiation from the sensor than they actually radiate, which the camera cannot detect directly.

To avoid these erroneous results, the threshold values of the color scale should be adjusted manually so that relevant areas can be better differentiated. Unfortunately, unlike more expensive models, the Thor002 does not allow you to assign different emission levels to certain areas of the image or even to specifically exclude conspicuous areas from the measured value, which reduces the informative value of the average and extreme values displayed. There is also no automatic segmentation or material database for emission correction.

Misinterpretation of reflective objects – gutter example

In this thermographic image, the gutter on the right stands out with a conspicuously low temperature value of supposedly -3.9 °C. However, a closer look at the scene and the physical conditions quickly reveals that this value is physically implausible. The measurement was taken on June 24 at a battery charge level of 96 %, which confirms that the camera was fully operational. The ambient temperature at this time was clearly in the plus range, so that frost or actual hypothermia can be ruled out.

The cause of this anomaly is the very low emissivity of metallic surfaces. The metal gutter has a considerably lower emissivity than the rendered façade, which makes it appear systematically colder in the thermal representation. Thor002 calculates the temperature from the amount of radiated heat and assumes a uniform emissivity for the entire image. At ε = 0.93, this is quite typical for building materials such as plaster, but far too high for the slightly weathered metal.

As the camera does not offer the option of storing different emissivities for individual surfaces, this inevitably results in a highly distorted image for reflective components such as rain gutters or window frames. Such errors can only be partially corrected in post-processing on the PC if the material and structure are known. Alternatively, targeted material markings, for example with insulating tape of known emissivity, can enable a more realistic measurement on site. Otherwise, there is a risk of misinterpreting metallic objects as so-called cold bridges, although this is merely a metrological effect. However, we saw on the previous page that it would be possible to limit the temperature ranges manually. But who would go to all this trouble when it is easier to do it afterwards on a PC?

Subsequent correction options on the PC – post-processing in ThermalSmart OS

The supplied PC software, which allows subsequent analysis of the stored IR images, provides a remedy here. In this environment, emission values for individual points or entire image areas can be subsequently adjusted and measurement errors corrected. This applies in particular if you know the structure and material of the scene or have taken reference images in parallel under controlled conditions. In this way, outliers on highly reflective or very cool surfaces, such as metal window frames or damp façade areas, can be classified much better. This opens up the possibility of raising the results to a higher level of analysis even with a low-cost camera such as the Thor002, provided you are prepared to continue working with the recorded data in post-processing. The combination of direct control on site and careful evaluation on the PC is therefore not only recommended, but even absolutely necessary for critical issues. Let’s return to the negative example and load this image directly from the camera onto the PC (memory card or via USB cable and camera memory).

The ThermalSmart OS software supplied makes it possible to make specific corrections to infrared images at a later date, which are not possible in the camera itself. This is particularly evident in the example of measuring point P2 on a reflective metal channel. In the initial image, a temperature value of -0.7 °C was determined here, although the real environmental context suggests a significantly higher temperature. This is due to the systematic underestimation of the radiated energy by the camera in the case of shiny and metallic surfaces. Individual emission values, reflected ambient temperature, atmospheric attenuation and the distance to the object can be entered manually using the “Region configuration” function in the software.

After correcting the emissivity to a value of 0.88, which is typical for heavily oxidized metal, there is already a visible adjustment of the temperature value. In addition, the correctly entered distance to the measuring point has a direct effect on the calculation accuracy, as the camera’s internal bolometer matrix performs distance compensation, taking atmospheric losses and radiation components into account.

How the software works in offline and online mode

The software offers two central operating modes. In offline mode, images and videos that have already been saved can be analyzed and post-processed. This includes setting measuring points, defining temperature zones, changing emission parameters, exporting to various formats and generating reports. Offline mode works completely locally, without an active connection to the camera. In online mode, on the other hand, there is a real-time connection to the Thor002, provided it is connected via USB or network. The live stream can be evaluated directly on the computer, including recording and remote control. This is particularly useful for stationary measurement setups or analyzing hard-to-reach areas. Alarm limits, recording series and automatic report generation can also be controlled via this.

The software is a useful addition to the camera and enables a more precise evaluation of the measurement results. Especially when working with different surface materials, the targeted adjustment of emissivity and environmental influences is essential in order to obtain realistic temperature values. The combination of simple operation and targeted post-processing means that the Thor002 can also be used for more technically demanding applications, provided the user is familiar with the basic principles of physics.

Comparison of accuracy on the black body radiator

The two thermal images show the measurement of a so-called “black body”, i.e. a temperature-stable, slightly spherically curved and matt black coated surface with an almost ideal emissivity of 1.0. The target temperature was deliberately set to 59 °C in order to avoid any internal compensation routines or “benchmark detection” of the camera software, which can occasionally occur with round values such as 60 °C. Such effects can occur particularly with low-cost cameras. Such effects can be observed above all with cheaper cameras, which could fall back internally on rounded or interpolated target value models for striking temperature values, which in practice is tantamount to a deliberate falsification of results.

The significantly higher-priced Optris Pi640, a precise industrial device with a resolution of 640 × 480 pixels and very low error deviation, served as a reference. This displayed a temperature of 58.8 °C in the center of the radiator. This corresponds to a deviation of just 0.2 Kelvin, which is within the specified tolerance limit.

The Thor002, on the other hand, whose sensor offers a significantly lower resolution and which is also in the lower price segment, reported a value of 57.2 °C for the same area. The deviation from the real value is therefore around 1.8 K. In view of the retail price of well under 400 euros, this is a really acceptable value, especially when you consider that the Thor002 was not originally developed for high-precision scientific measurements.

Measured against its price-performance ratio, the Thor002 therefore delivers a thoroughly respectable result. Under controlled conditions and ideal emissivity, the measurement accuracy of the camera is very close to that of higher-class devices, albeit with a slight systematic offset. For applications in construction and building services, in maintenance or for thermal process controls in the IT environment, however, this behavior can be classified as sufficiently accurate. However, the Thor002 is not absolutely calibrated or adjusted before series delivery, as is the case with the Optris, but only undergoes a simple rough calibration. If you want to carry out really precise reference measurements, you should always be aware of the possible deviations and, if necessary, check with known reference bodies.