SWIR Cameras Without Sensor Cooling

Advantages of SWIR cameras without TEC sensors

• The camera housing can feature a smaller design because no internal cooling is required

• Significantly lower price

• No fans that generate air turbulence, which is particularly advantageous in clean rooms

• No vibrations due to fans, which avoids measurement inaccuracies

• Lower power consumption, as no thermoelectric cooling modules or fans are required

• Operation via just one cable is possible, also with USB

Sensor temperatures and signal-to-noise ratio

The relevance of the individual aspects of image quality depend heavily on the requirements of the application. For some applications it is essential that the noise is minimal, for others it is a matter of the greatest possible dynamic range. And still others need to identify the smallest structures. This is why there is not just one parameter that represents image quality, but several. Because only the influence of temperature and therefore the dark current noise are relevant in this case, the SNR is a suitable indicator to quantify the changes in image quality. As the SNR only considers homogeneous surfaces, structures and the edge profile must also be taken into account.

Camera setups with representative sensor temperatures

In order to compare the image quality, we use four different camera setups with typical sensor temperatures at maximum frame rate:

• Uncooled, unmounted camera with approx. 60 °C sensor temperature

• Uncooled, unmounted camera with approx. 45 °C sensor temperature

• Camera actively cooled with compressed air cooler with approx. 33 °C sensor temperature

• Camera with TEC sensor with approx. 15 °C sensor temperature

Exposure times under 10 milliseconds

Extensive series of measurements show that TEC sensors offer no advantage in terms of image quality at exposure times of less than 10 ms. This exposure time corresponds to a frame rate of 100 fps with continuous image acquisition. SenSWIR technology already minimizes the typical problems of conventional InGaAs sensors here.

Even when comparing very different sensor temperatures (15 °C vs. 60 °C) with an exposure time of 10 ms, no particular deviation in the image quality of TEC and uncooled SenSWIR sensors can be determined. The SNR shows no significant difference. The same result can be seen in the qualitative view of the structures or the edge profile.

Sensor cooling therefore does not provide a decisive advantage - the influence of the dark current is negligible in this exposure time range. Around 80 % of all SWIR applications use such short exposure times.

Exposure times above 10 milliseconds

With longer exposure times, the sensor temperature plays a greater role. As the sensor temperature rises, the influence of the dark current increases and the SNR decreases, causing the image quality to deteriorate. The main cause is defective pixels, the so-called "hotspots", which can be recognized as white dots in the image. The higher the sensor temperature and the longer the exposure time, the more hotspots occur (Fixed Pattern Noise - FPN). At the same time, the higher dark current also increases the so-called dark current shot noise, which manifests itself as random noise (RN).

A TEC sensor generally improves the image quality for exposure times above 10 ms. However, the disadvantages of a sensor without TEC can be compensated by an external cooling solution and dedicated firmware depending on the respective application setup and the image quality requirements. It is therefore worth checking whether the image quality of an uncooled SenSWIR sensor is sufficient in the application to benefit from its advantages.

Pixel Correction Beyond for the reduction of pixel errors

Our unique Pixel Correction Beyond algorithm was developed to eliminate the pixel errors commonly found in industrial SWIR imaging. By incorporating the image content and flexibly adjusting the correction strength, significantly better SWIR image results are achieved.

Reduction of the blooming effect

Under extreme conditions such as a sensor temperature of 60 °C and an exposure time greater than 100 ms, so-called "blooming hotspots" can occur. These are clusters of hotspots that occur when a hotspot radiates onto other pixels. In this case, the "Blooming Reduction", a SenSWIR-specific feature from Sony, can be used. This minimizes the occurrence of blooming hotspots, although this is only successful in around 50 % of all cases. Whether or not blooming hotspots are a problem must be determined for each individual application.

PGI for noise reduction and image sharpness

Our tried-and-tested PGI feature set offers another way to optimize image quality. Denoising reduces random noise without affecting the edges. To compensate for blurring of the image, Sharpening can be used to accentuate the edges and additionally optimize the sharpness of the image. This makes it possible to individually adapt the image quality to the inspection task.

Pixel Correction Beyond and Denoising

What effect does the combination of Pixel Correction Beyond and Denoising have on the edge profile? The noise is reduced while the structures in the image are retained. Unlike with simple noise reduction algorithms, the gradient and amplitude between the high and low intensity (gray) value levels are retained. With stronger noise reduction, it should be noted that blurring in the image can increase. Here it is important to carefully balance the parameters for noise reduction and sharpening.

TECless cameras offer many advantages

The general assumption that TEC sensors always deliver better image quality in SWIR imaging is incorrect. Rather, it depends on the specific application. For example, TEC-SenSWIR sensors offer no advantage in terms of image quality compared to uncooled SenSWIR sensors at exposure times of less than 10 milliseconds. On the other hand, there are the many disadvantages of cameras with TEC sensors. For longer exposure times, our combined solution of cooling accessories and camera features offers a cost-effective alternative to the high-priced cameras with TEC sensors. Depending on the application, this solution achieves the image quality of cameras with TEC sensors. It is therefore worth taking a closer look to find the optimum solution for your individual application.