Co-located sensors in a sorting facility

1 . A system, comprising:

a first sensor associated with a first field of view, and wherein the first sensor is located in a sorting facility;

a second sensor associated with a second field of view, wherein the second field of view is adjacent to or overlaps with the first field of view, and wherein the second sensor is located in the sorting facility, wherein a combination of the first field of view and the second field of view covers at least a width of a conveyor device; and

one or more processors configured to:

receive a first image including a first set of objects from the first sensor, wherein the first set of objects is located on the conveyor device;

receive a second image including a second set of objects from the second sensor, wherein the second set of objects is located on the conveyor device;

generate a master image by combining the first image and the second image, wherein the first image shows a target object with an occluded portion, wherein the occluded portion of the target object is visible within the master image;

determine a set of attributes associated with the target object based at least in part on using a machine learning model to evaluate the master image, wherein the machine learning model has been trained on master images that have been stitched together from images captured by individual sensors;

determine a sorting operation to be performed on the target object based at least in part on the set of attributes associated with the target object, wherein the sorting operation to be executed by a set of sorting devices that is located downstream from the first sensor and the second sensor in the sorting facility;

obtain, by an audit sensor, a third image depicting an execution of the sorting operation on the target object to divert the target object towards a desired destination, wherein the audit sensor comprises a third sensor located downstream from the first sensor and the second sensor in the sorting facility;

determine an audit result based on whether the sorting operation was successful or not based on the third image by programmatically determining whether the target object appears to be successfully directed towards the desired destination in the third image;

determine a capture success rate associated with the set of sorting devices based at least in part on the audit result; and

determine whether to retrain the machine learning model used to detect the set of attributes associated with the target object based at least in part on the capture success rate.

2 . The system of claim 1 , wherein the first sensor and the second sensor are vision sensors.

3 . The system of claim 1 , wherein the first image is captured by the first sensor at a first time and the second image is captured by the second sensor at a second time, and wherein the first time and the second time are within a threshold window of time.

4 . The system of claim 1 , wherein the first image and the second image both show the target object.

5 . The system of claim 1 , wherein the first sensor and the second sensor are placed over a conveyor belt.

6 . The system of claim 1 , wherein the first sensor is placed at a first angle relative to the target object and the second sensor is placed at a second angle relative to the target object, and wherein the first angle is different from the second angle.

7 . The system of claim 1 , wherein the set of sorting devices includes a first sorting device that uses a first sorting mechanism type and a second sorting device that uses a second sorting mechanism type.

8 . The system of claim 7 , wherein the first sorting mechanism type and the second sorting mechanism type are the same.

9 . The system of claim 8 , wherein the first sorting mechanism type and the second sorting mechanism type are different.

10 . The system of claim 1 , wherein the set of sorting devices includes a first sorting device and a second sorting device in a dual cut sorting arrangement.

11 . The system of claim 1 , wherein to determine the sorting operation comprises to determine a surface of the target object with which a sorting device should engage during the sorting operation.

12 . The system of claim 1 , wherein to evaluate the first image and the second image comprises to:

perform cross-frame comparison between the first image and the second image to improve segmentation of the target object; and

wherein to determine the set of attributes associated with the target object is further based at least in part on the cross-frame comparison.

13 . The system of claim 1 , wherein the master image comprises a three-dimensional image.

14 . The system of claim 1 , wherein to generate the master image comprises to stitch together the first image and the second image.

15 . The system of claim 1 , wherein to retrain the machine learning model comprises to retrain the machine learning model with new master images that are annotated with object locations and attributes.

16 . The system of claim 1 , wherein the one or more processors are further configured to reconfigure a sorting parameter associated with a sorting device that had performed the sorting operation based at least in part on the capture success rate.