Depth information processing method, apparatus, and storage medium

1. A depth information processing method, comprising:

determining intermediate depth information of a target scene according to sparse depth information of the target scene by a sub-model unit in a depth information supplementing model; and

using intermediate depth information determined by a tail sub-model unit in the depth information supplementing model as dense depth information of the target scene;

wherein the determining the intermediate depth information of the target scene according to the sparse depth information of the target scene by the sub-model unit in the depth information supplementing model comprises:

processing, by an hourglass network layer of the sub-model unit in the depth information supplementing model, inputted depth information to obtain supplementary depth information; and

superimposing the supplementary depth information with the inputted depth information of the sub-model unit by an accumulation layer of the sub-model unit to obtain the intermediate depth information determined by the sub-model unit.

2. The depth information processing method of claim 1 , wherein the superimposing the supplementary depth information with the inputted depth information of the sub-model unit comprises:

determining a matching relationship between supplementary pixels in the supplementary depth information and sparse pixels in the inputted depth information by the accumulation layer of the sub-model unit; and

superimposing supplementary depth data of the supplementary pixels with sparse depth data of matched sparse pixels.

3. The depth information processing method of claim 1 , further comprising:

determining depth feature information according to pixel information of the target scene by the sub-model unit in the depth information supplementing model; and

adjusting the intermediate depth information determined by the sub-model unit according to the depth feature information.

4. The depth information processing method of claim 1 , further comprising:

acquiring standard sparse depth information and standard dense depth information of a standard scene; and

performing training according to the standard sparse depth information and the standard dense depth information to obtain the depth information supplementing model.

5. The depth information processing method of claim 4 , wherein the acquiring the standard sparse depth information and the standard dense depth information of the standard scene comprises:

performing image collection on the standard scene using a depth sensor to obtain the standard dense depth information; and

sampling the standard dense depth information to generate the standard sparse depth information.

6. The depth information processing method of claim 4 , wherein the acquiring the standard sparse depth information and the standard dense depth information of the standard scene comprises:

performing video collection on the standard scene using a radar device to obtain a plurality of consecutive frames of sparse depth information;

projecting the plurality of consecutive frames of sparse depth information to generate the standard dense depth information; and

acquiring sparse depth information matching the standard dense depth information from the plurality of consecutive frames of sparse depth information and determining the acquired sparse depth information as the standard sparse depth information.

7. An electronic apparatus, comprising:

at least one processor; and

a memory communicatively connected to the at least one processor;

wherein the memory stores an instruction executable by the at least one processor to enable the at least one processor to perform:

determining intermediate depth information of a target scene according to sparse depth information of the target scene by a sub-model unit in a depth information supplementing model; and

using intermediate depth information determined by a tail sub-model unit in the depth information supplementing model as dense depth information of the target scene;

wherein the determining the intermediate depth information of the target scene according to the sparse depth information of the target scene by the sub-model unit in the depth information supplementing model comprises:

processing, by an hourglass network layer of the sub-model unit in the depth information supplementing model, inputted depth information to obtain supplementary depth information; and

superimposing the supplementary depth information with the inputted depth information of the sub-model unit by an accumulation layer of the sub-model unit to obtain the intermediate depth information determined by the sub-model unit.

8. The electronic apparatus according to claim 7 , wherein the superimposing the supplementary depth information with the inputted depth information of the sub-model unit comprises:

determining a matching relationship between supplementary pixels in the supplementary depth information and sparse pixels in the inputted depth information by the accumulation layer of the sub-model unit; and

superimposing supplementary depth data of the supplementary pixels with sparse depth data of matched sparse pixels.

9. The electronic apparatus according to claim 7 , wherein the instruction, executed by the at least one processor, further enables the at least one processor to perform:

determining depth feature information according to pixel information of the target scene by the sub-model unit in the depth information supplementing model; and

adjusting the intermediate depth information determined by the sub-model unit according to the depth feature information.

10. The electronic apparatus according to claim 7 , wherein the instruction, executed by the at least one processor, further enables the at least one processor to perform:

acquiring standard sparse depth information and standard dense depth information of a standard scene; and

performing training according to the standard sparse depth information and the standard dense depth information to obtain the depth information supplementing model.

11. The electronic apparatus according to claim 10 , wherein the acquiring the standard sparse depth information and the standard dense depth information of the standard scene comprises:

performing image collection on the standard scene using a depth sensor to obtain the standard dense depth information; and

sampling the standard dense depth information to generate the standard sparse depth information.

12. The electronic apparatus according to claim 10 , wherein the acquiring the standard sparse depth information and the standard dense depth information of the standard scene comprises:

performing video collection on the standard scene using a radar device to obtain a plurality of consecutive frames of sparse depth information;

projecting the plurality of consecutive frames of sparse depth information to generate the standard dense depth information; and

acquiring sparse depth information matching the standard dense depth information from the plurality of consecutive frames of sparse depth information and determining the acquired sparse depth information as the standard sparse depth information.

13. A non-transitory computer-readable storage medium, storing a computer instruction for causing a computer to perform the following steps:

determining intermediate depth information of a target scene according to sparse depth information of the target scene by a sub-model unit in a depth information supplementing model; and

using intermediate depth information determined by a tail sub-model unit in the depth information supplementing model as dense depth information of the target scene;

wherein the determining the intermediate depth information of the target scene according to the sparse depth information of the target scene by the sub-model unit in the depth information supplementing model comprises:

processing, by an hourglass network layer of the sub-model unit in the depth information supplementing model, inputted depth information to obtain supplementary depth information; and

superimposing the supplementary depth information with the inputted depth information of the sub-model unit by an accumulation layer of the sub-model unit to obtain the intermediate depth information determined by the sub-model unit.