Thermal mapping

1. A method for thermal mapping for additive manufacturing by an electronic device, comprising:

obtaining a two-dimensional contone map, at a voxel-level resolution, for additive manufacturing;

obtaining a first thermal image at a first resolution; and

calculating, using a neural network, a second thermal image at a second resolution based on the two-dimensional contone map and the first thermal image, wherein the second resolution is at the voxel-level resolution and greater than the first resolution.

2. The method of claim 1 , further comprising concatenating first data and second data, wherein the first data is based on the first thermal image and the second data corresponds to the map or a plurality of maps.

3. The method of claim 1 , wherein the neural network is a recurrent neural network comprising a convolutional long short-term memory network or multiple stacked convolutional long short-term memory networks.

4. The method of claim 1 , further comprising increasing the first resolution of the first thermal image to produce an upscaled first thermal image.

5. The method of claim 4 , further comprising: encoding the upscaled first thermal image to produce first data that is provided to the neural network; encoding the map to produce second data that is provided to the neural network; and decoding an output of the neural network to produce the second thermal image.

6. The method of claim 5 , wherein the encoding the first thermal image is performed with a first neural network, the encoding the map is performed with a second neural network, and the decoding the output is performed with a third neural network.

7. The method of claim 1 , wherein the second resolution is greater than the first resolution by a factor of at least 8.

8. The method of claim 1 , wherein the first thermal image corresponds to a first layer and the second thermal image corresponds to the first layer, wherein the second thermal image is an enhanced voxel-level thermal image of the first layer.

9. The method of claim 1 , wherein the first thermal image corresponds to a first layer and the second thermal image corresponds to a second layer that is subsequent to the first layer, wherein the second thermal image is a predicted voxel-level thermal image calculated before forming the second layer, and wherein the method further comprises performing feedback control based on the second thermal image.

10. A three-dimensional (3D) printing device, comprising:

a print head to print a fusing agent based on a two-dimensional fusing contone map at a voxel-level resolution and to print a detailing agent based on a detailing contone map;

a thermal projector;

a thermal sensor; and

a controller, wherein the controller is to:

receive a captured thermal image of a first layer from the thermal sensor, wherein the captured thermal image is at a first resolution; and

predict, using a neural network based on the two-dimensional fusing contone map,

the detailing contone map, and the captured thermal image,

a predicted thermal image of a second layer that is subsequent to the first layer, wherein the predicted thermal image is at a second resolution that is at the voxel-level resolution and greater than the first resolution.

11. The 3D printing device of claim 10 , wherein the controller is to:

upscale the captured thermal image to produce an upscaled thermal image;

encode, using a first convolutional neural network, the upscaled thermal image to produce first data;

encode, using a second convolutional neural network, the fusing contone map and the detailing contone map to produce second data; and

concatenate the first data with the second data to produce concatenated data, wherein the concatenated data is input to the neural network.

12. The 3D printing device of claim 11 , wherein the controller is to decode, using a third convolutional neural network, an output of the neural network to produce the predicted thermal image.

13. A non-transitory tangible computer-readable medium storing executable code, comprising:

code to cause a processor to obtain a two-dimensional fusing contone map, at a voxel-level resolution, for additive manufacturing, and a detailing contone map corresponding to a layer;

code to cause the processor to obtain a captured thermal image corresponding to the layer; and

code to cause the processor to enhance the captured thermal image using a neural network based on the two-dimensional fusing contone map, the detailing contone map, and the captured thermal image to produce an enhanced thermal image with an increased resolution relative to a resolution of the captured thermal image, whereby the increased resolution is at the voxel-level resolution.

14. The computer-readable medium of claim 13 , further comprising code to train the neural network using a set of training thermal images with a second resolution that is greater than the resolution of the captured thermal image.