Shopping Cart

Frequently Asked Questions: Thermal Imaging for Unmanned Aerial Systems (Part 1)

16 July, 2021 | Resources

Featured image for Frequently Asked Questions: Thermal Imaging for Unmanned Aerial Systems (Part 1)

There are many questions asked around thermal payloads for drones. Teledyne FLIR has answered some of these questions below.

Q: What is the difference between 30Hz and 90Hz?

A: Video cameras output a series of “still” images at a regular rate, a rate given in Hertz (Hz) or in frames per second (fps). FLIR UAS cameras with 25 to 60 fps video put out that number of unique frames each second. To make transportation of cameras across borders easier, a class of “<9 Hz” cameras is available. These cameras send out video frames at a normal rate, but these frames are duplicated for short intervals. This results in an effective frame rate of fewer than 9 fps. Please contact an export specialist before transporting any thermal camera from one country to another.

Q: Can I export or travel internationally with my thermal camera?

A: There are international regulations regarding the transportation and transfer of all thermal cameras. Export laws allow thermal cameras with frame rates less than 9 frames per second (fps) to be moved more freely across borders than 60 or 30 fps cameras. Contact an export specialist for details.

Q: What is the difference between 160x120, 336x256, and 640x512 resolution cameras?

A: Each FLIR UAS thermal camera has an imaging device known as a focal plane which converts the target image to pixels. The sizes 160 × 120, 336 × 256, and 640 × 512 are examples of the resolution options available for these cameras. The first value represents the horizontal pixel count, and the second number represents the vertical pixel count. At a quick glance, many people assume the 640 is twice the resolution at a 336 but in fact, the total pixel count is 3.8x more. 336 × 256 delivers 86,016 total pixels and 640 × 512 delivers 327,680 pixels.

Q: What is the field of view?

A: Field of view, also referred to as FOV, is the degree of visibility the camera lens delivers to the sensor. For example, each pixel in the 640 × 512, 13 mm configuration (45 × 37-degree FOV) will represent an angle of 0.07 degrees per pixel, meaning that at 100ft, each pixel is imaging 1.57 inches.  From a less technical perspective, FOV equates to the observable area that can be imaged with the lens. The graphics below show how various FOVs in 640 and 256 cameras correspond to the observable area when looking straight down. The calculation is linear, so an elevation of 200 ft would double the value., while an elevation of 50 ft would be half of the value.

Q: What is the best lens for my operation?

A: There are two major considerations in determining the best lens and resolution for a given application. First, the field of view (FOV) is the approximate angle of the observable image.   Figure 2 shows the observable area by lens model at 100 ft AGL assuming you are looking straight down.

The second factor is the angle that each pixel represents, known as the iFOV. Knowing the angle of each pixel helps you calculate the number of pixels that would appear for various targets at various distances. For temperature measurement, an object should appear as at least a 5 × 5-pixel grid but 10 × 10 or more will yield even better results. A lens should be chosen so that the FOV is wide enough to find what you are looking for, and that the iFOV will allow enough “pixels on target” at the desired flight altitude. Figure 2 demonstrates spot size at 100 ft AGL when looking straight down. Globally, the 640 × 512, 13 mm option is the #1 selling model and performs most tasks very well.

Q: What should I buy?

A: Camera systems are often available as a package with a compatible drone. This is the easiest way to ensure compatibility. There are gimbal packages available that target specific cameras on specific drones or drone families. These may be available from the vehicle manufacturer or third parties, such as Gremsy. Documentation for each gimbal will identify which cameras and vehicles they support.

To integrate a general-purpose sUAS camera to your existing drone, several factors need to be considered.

Weight and centre of gravity (CG): The drone must be able to lift the camera and still maintain a useful flight time. Some provisions need to be made to support the camera on the vehicle and maintain balance. FLIR UAS cameras usually have a ¼-20 threaded hole on either the top, bottom, or both sides of the camera. These holes are the same size and thread pattern found on consumer cameras so they can be used on tripods.

Electrical supply: Most FLIR cameras are designed to draw power directly from the vehicle battery, while others draw regulated 5 volts from a battery elimination circuit (BEC). The voltage and current requirements of each camera can be found in the datasheet or user guide for the camera.

Video: HDMI or composite video outputs may be available from the camera. If real-time video is needed for your mission, a compatible video downlink system must be provided.

Camera and gimbal control: Various features of the camera and any associated gimbal will need to be controlled. Often, this is done using servo PWM signals. These are three-pin outputs from flight controller or remote-control receiver that are used to control motor controllers, servo actuators, and other vehicle components.

GPS: Some cameras have built-in GPS receivers for geotagging image files. These receivers may need to be connected to external antennas, which are most often placed on top of the vehicle for the best possible satellite reception. Some cameras do not have internal GPS receivers, and if image files are to be geotagged, information from the GPS on the vehicle will need to be sent to the camera. Typically, this will be done using MAVLink. The engineering datasheet or user guide will list the available interfaces and their capabilities.

Q: Can I detect methane gas leaks with my thermal camera?

A: Gas-finding IR cameras work by detecting the absorption of light or heat energy by gasses. This absorption is typically weak and in a narrow spectrum, requiring specialized optical filters and high-quality sensors. General-purpose thermal will not reliably detect these gasses. Gas-finding cameras are specially designed with narrow optical filters and typically contain a cryogenic cooler, which draws additional power, and adds weight.

Q: How high can I fly with my thermal camera?

A: The maximum allowable altitude can be found under the environmental specifications on the product datasheet. While this is typically 38,000 feet MSL for most products, UAS vehicles must be operated according to applicable laws. Also, camera resolution imposes practical limitations on useful altitudes.  At an altitude of 400 feet, one pixel on a Duo Pro R camera with a 13 mm lens would represent an area more than 6 inches across. For most thermal imaging applications, the target would have to be much closer to provide enough pixels on target for the image to be useful, or a camera with a different lens must be used. For example, at the same 400-foot altitude and a 25 mm lens, one pixel would represent an area ~3.4 inches across. For temperature measurement, an object should appear as at least a 5 × 5-pixel object in the image.  10 × 10 will yield even better results.

Q: How far can I see based on my lens choice?

A: There is no practical limit to how far a thermal camera can see through a clear line of sight but understanding what you are seeing is important.   The moon is often visible when not obscured by clouds (water vapor absorbs IR energy). Seen from Earth, the moon subtends an angle of about 0.5 degrees, so on a 25-degree FOV 640 resolution camera, the moon would appear as about a 12-pixel wide circle.

The normal limit for the target distance of an IR camera is the size in pixels that the object will appear in the image and the number of pixels that are required to identify or extract the required information about the target. For example, as shown in the image, do you want to Detect, Recognize, or Identify the person in the image?   The number of pixels you place on an item allows you to make the determination.   For temperature measurement, an object should appear as at least a 5 × 5-pixel object in the image. 10 × 10 will yield even better results.

Q: Can I create 3D orthomosaics with my thermal datasets?

A: FLIR does not currently offer tools for thermal orthomosaics. For mapping operations, there are third party thermal orthomosaic services experience in processing images from thermal cameras.   Currently, the ability to retain pixel-level radiometric temperature data is lost in the process of creating the ortho.

Q: What software is best for my thermal data set or operation?

A: Different cameras offer different file types and characteristics. Also, different applications have different requirements. Higher-resolution (320 × 256 or greater) thermal cameras can achieve a better level of detail by saving images in the -R.JPEG format and processing them with the free FLIR Tools application. This combination allows radiometric parameters to be adjusted post-flight and provides a simple framework for generating reports. CSV files of pixel values can also be created for additional processing.  FLIR Thermal Studio offers batch processing for jobs that require many files. 

Source: FLIR