Object imaging and detection systems and methods
A method including selecting image data of a mechanical mud separation machines (“MMSM”) to detect objects in an object flow and other operational conditions at the MMSM. The image data may be processed by a Deep Neural Network to identify objects in the object flow, operational parameters of the MMSM, and environmental conditions. Additional image data may be selected for additional processing based on the results of the analysis.
1. A computer implemented method comprising:
receiving a field of view comprising image data from an imaging device capturing images of a mechanical mud separation machine (“MMSM”);
selecting, from the image data, a region of interest (“ROI”), wherein selecting the ROI includes:
identifying a ledge of an MMSM within the image data, the ledge being a location where objects in an object flow begin to free fall;
setting a top edge of the ROI at the ledge;
identifying a frames per second of an imaging device capturing the image data;
calculating a vertical length of the ROI based on the frames per second of the imaging device and the ledge;
associating the image data with a timestamp that corresponds to a time at which the image data was captured by the imaging device.
2. The method of claim 1 , further comprising: preprocessing the ROI to form a preprocessed ROI.
3. The method of claim 2 , wherein the preprocessing comprises at least one selected from the group consisting of rotating the image data such that a ledge of the MMSM is substantially horizontally positioned, cropping the ROI to a specified pixel size, brightness equalization, and white balancing the image.
4. The method of claim 1 , further comprising analyzing, using a DNN, the preprocessed ROI to identify a plurality of objects in an object flow;
identifying, using a DNN, a first object in an object flow; and
associating the object with a first-time stamp.
5. The method of claim 1 , wherein identifying the objects comprises at least one of estimating the object's size, shape, or color.
6. The method of claim 1 , further comprising:
analyzing, using a DNN, the ROI to identify a plurality of objects in an object flow;
classifying each of the plurality of objects to form a plurality of classified objects;
aggregating each of the classified objects into one or more groupings;
determining the number of classified objects within each of the one or more groupings;
comparing the number of objects within each of the one or more groupings to a threshold;
based on the comparison, determining that an anomaly is present; and
based on determining that an anomaly is present, sending a signal to begin capturing additional image data to an additional imaging device.
7. The method of claim 6 , wherein the threshold is determined at least in part by signal data.
8. The method of claim 7 , wherein the signal data comprises at least one selected from the group consisting of: a motor current, a temperature reading, a light meter, and a wind gage.
9. The method of claim 6 , further comprising:
based on determining that an anomaly is present, sending a signal to the imaging device to change at least one image device setting selected from the group consisting of: a shutter speed, a frame rate per second, an aperture setting, and a resolution setting.
10. The method of claim 6 , wherein the classified objects comprise cuttings and updating a cuttings transport model based on the classification of the cuttings.
11. The method of claim 6 , wherein the classifying operation occurs at a different location from the receiving a field of view comprising image data operation.
12. A non-transitory computer-readable storage device storing instructions that, when executed, perform the method of
receiving a field of view comprising image data from an imaging device capturing images of a mechanical mud separation machine (“MMSM”);
selecting, from the image data, a region of interest (“ROI”), wherein selecting the ROI includes:
identifying a ledge of an MMSM within the image data, the ledge being a location where objects in an object flow begin to free fall;
setting a top edge of the ROI at the ledge;
identifying a frames per second of an imaging device providing the image data;
calculating a vertical length of the ROI based on the frames per second of the imaging device and the ledge;
associating the image data with a timestamp that corresponds to a time at which the image data was captured by the imaging device.
13. The non-transitory computer-readable storage device of claim 12 , wherein the method further comprises:
analyzing, using a DNN, the ROI to identify a plurality of objects in an object flow;
classifying each of the plurality of objects to form a plurality of classified objects;
aggregating each of the classified objects into one or more groupings;
determining the number of classified objects within each of the one or more groupings;
comparing the number of objects within each of the one or more groupings to a threshold;
based on the comparison, determining that an anomaly is present; and
based on determining that an anomaly is present, sending a signal to begin capturing additional image data to an additional imaging device.