**Luxonis continues to add more and more sensors to their platform** We are thrilled to announce that our OAK-Thermal (OAK-T) is now **available for pre-order in our Early Access store ([here](https://shop.luxonis.com/collections/early-access/products/oak-t))**. Rounds of prototypes have already been completed and we’ll be shipping out our first mass production deliveries in the beginning of April. If you pre-order now then you’ll be first in line to receive our first production units. The decision to add a thermal sensor to OAK was made by our customers who frequently requested the ability to fuse thermal data with high resolution RGB all on a compact edge device. The OAK-T opens up all kinds of new potential applications such as detecting leaks that the vision based systems (and humans) are unable to detect. Additionally, the OAK-T will be useful for detecting humans & animals, monitoring equipment, detecting fires, and much more! Thermal cameras detect infrared radiation, which is emitted by all objects based on their temperature. We use a special IR-capable lens to focus the infrared light emitted by all of the objects/scene in view, which is then scanned by an array of infrared-detector elements ("pixels") in the camera. Here are some key technical details: * **Thermal sensitivity:** 50mK @ 25°C and 25Hz * **Measurement range:** -15°C - 150°C * **Thermal Accuracy:** ±2°C or ±2% (whichever is larger) * **Sensor type:** Uncooled VOx detector * **Spectral range:** 8-14µm [upl-image-preview url=https://discuss.luxonis.com/assets/files/2024-01-19/1705675593-139359-thermal-details.webp] And now, a little more background on why Luxonis keeps adding so many new sensors to their platform. At Luxonis, we are sensor agnostic. What this means in practice is that we are continually adding new and improved sensors to our platform. We are not married to any one type of sensor or approach as we work towards our ultimate goal which is to enable robotic and automation systems to exceed human level perception at every task. We are able to support so many different sensors because we self-perform all of our engineering down to the bare metal up to the firmware drivers and software. For example, we already offer dozens of imagers that are fully supported by our DepthAI API ([here](https://docs.luxonis.com/projects/hardware/en/latest/pages/articles/supported_sensors/)) and these include a 32MP IMX582 option ([here](https://shop.luxonis.com/collections/oak-modules/products/oak-ffc-imx582)) along with a recently added IMX477-M12 ([here](https://shop.luxonis.com/collections/oak-modules/products/oak-ffc-imx477)). The relentless addition of new sensors is not limited to only cameras. We already support inertial measurement units (IMU), infrared illumination, and infrared laser dot projector. Recently, we added our first Time of Flight (ToF) that is also available for order now with our next deliveries scheduled for mid-February ([here](https://shop.luxonis.com/collections/early-access/products/oak-d-sr-poe)). Luxonis has also supported other sensor types for custom development projects. Stay tuned as we will be coming out with more new sensors this summer! If you have any ideas for more sensors that we should support please let us know at [support@luxonis.com](mailto:support@luxonis.com) or by posting on our [forum](https://discuss.luxonis.com/). You can pre-order our OAKs with new sensors now: - [Thermal](https://shop.luxonis.com/collections/early-access/products/oak-t) - [Time of Flight](https://shop.luxonis.com/collections/early-access/products/oak-d-sr-poe)

Please give us more specs on the IR imager particularly, what is its resolution? Thanks

From the specs, this camera is the answer to my prayers, except for one thing: Gimbal compatibility The connectors on the back look like they would seriously hinder gimbal movement, due to their position and the (possible) sturdiness of the cables. I really can't understand why you never seem to consider gimbal compatibility for your cameras. That's a real shame, as your products are great and, more important, unique on the market. Could there be an USB version so you can attach a lightweight flat cable? If not, what is the name of the connectors on the back, so one can fashion custom adapter boards?

We are developing our Series 4 version of the Thermal camera and that new version will come with the option to connect via POE or USB. The connectors on the back are M12 and M8.

Thanks for the info. Both connectors look like M series connectors from Amphenol (https://www.amphenolltw.com/product-info/M-Series/M-Series.M8/M8-08BFFM-SR7B10.html, https://www.amphenolltw.com/product-info/M-Series/M-Series.M12%28X%29/M12X-08PFFT-EE0003.html). Is that correct? How is the camera connected to the computer? It seems the M12 plug goes to a PoE cable, making this camera similar to your existing PoE models? What would the M8 socket be for?

Yes, the connectors are indeed M series connectors from Amphenol. The M8 aux connector is an 8pin female A-coded connector, and the M12 ethernet is an 8pin female X-coded connector. The M12 plug does indeed go to a PoE cable, making this camera similar to existing PoE models. The M8 socket is used for auxiliary connections. If the M8 connector is not used, the cameras include an M8 connector cap for waterproofing. To connect the camera to your computer, you will need a PoE switch or injector to power the device. After powering the device, you should connect your computer to the same Local Area Network (LAN) as the PoE device. If you are using a DHCP server, you can check the connected devices on the DHCP server dashboard to see if the camera is connected and its IP. If there is no DHCP, the camera will fallback to a static IP. Once you know the IP of the camera and can ping it, you can run any code sample or depthai experiment as you would when connecting an OAK device with a USB-C cable.

Hi @"LianyuChu"#2242 , 1) Yes, you'll be able to use depthai API library to access both streams 2) Yes 3) Yep, we already have a demo below. It does inferencing on the thermal frame directly: https://github.com/luxonis/depthai-python/blob/thermal_rvc2_nnet/examples/NeuralNetwork/thermal_nnet.py (currently uses local path to NN, we'll upload it to our server soon) 4) Yes, correct

Return to blog overview

CategoryCategory

Ghost title

Comments (0)

Categories
Blog Releases
Forum Nav

Announcing our first OAK with a thermal sensor

LianyuChu

I am excited to see this product. I would like to obtain more info in order to know if it may meet our needs.

(1) Will OAK-T provide both video and thermal data like the rtsp streams?

(2) Will users access the thermal raw data at 256*192 resolution and 25 Hz?

(3) Is it possible for users to use their own model like Yolo8?

(4) It sounds like that OAK-T uses the Tiny-1C module from Infiray. Is it correct?

Thanks

erik

Hi @LianyuChu ,
1) Yes, you'll be able to use depthai API library to access both streams
2) Yes
3) Yep, we already have a demo below. It does inferencing on the thermal frame directly: luxonis/depthai-pythonblob/thermal_rvc2_nnet/examples/NeuralNetwork/thermal_nnet.py (currently uses local path to NN, we'll upload it to our server soon)
4) Yes, correct

LianyuChu

Erik:

I just pre-ordered one and can't wait to test it. I have one more question:
- when regular video is processed by OAK-SOM-Pro, what is the normal image size that can be used in the AI model? We are currently using Google TPU. We have to use an image size like 192*192 of Yolo8 model.

Thanks

jakaskerl

LianyuChu
Depends on the overall FPS you wish to have. I have seen models go to 1080x1080 or even more. Of course, bigger input size will mean the NN is more computationally expensive. So basically any image size will work, but at the end, it's a compromise.

Most basic size (used by many examples) is 300x300.

Thanks.

LianyuChu

Does the OAK-T provide depth info like other luxonis depth cameras? If so, OAK-T uses both visual camera and thermal camera info to get the depth estimate. Is it correct?

jakaskerl

Hi @LianyuChu
Oak-T does not provide depth since it does not feature stereo cameras (or ToF sensor).

RBJin

Do you have any plans to design the thermal sensor as a module similar to OAK-FFC ToF33D ?

erik

Hi @RBJin ,
Because the thermal sensor uses SPI it's not possible to connect it that way (via MIPI lines). We have developed (for internal use only) the board below to be able to connect it to the FFC-4P.

LianyuChu

I preordered one but found its release has been delayed for a few times. Now, it shows August. Is it possible to get the delta version of the product and we probably can help for testing?

Thanks

LianyuChu

I received the Oak Thermal. How to start using it? Where may I find a user manual or instruction?

BradleyDillon

@LianyuChu You can use our guide here to get started: https://docs.luxonis.com/hardware/platform/deploy/poe-deployment-guide/

Ghost title

Comments (0)

Categories

Forum Nav