Steel strip surface grinding image detection system and steel strip surface grinding image detection method

Abstract

A steel strip surface grinding image detection system includes a grinding image database and a processor. The grinding image database is configured to store at least one grinding image. The processor is communicatively connected to the grinding image database and configured to sequentially perform the following steps: accessing the grinding image database and obtaining the grinding images; then establishing a grinding area detection model based on the grinding images; locating the grinding area in the steel strip surface image using the grinding area detection model; and detecting the grinding area using the grinding area detection model to obtain a grinding area detection result.

Claims

1. A steel strip surface grinding image detection system, comprising: a grinding image database configured to store at least one grinding image; and a processor communicatively connected to the grinding image database and configured to sequentially perform the following steps: accessing the grinding image database and obtaining the at least one grinding image, then establishing a grinding area detection model based on the at least one grinding image; locating a grinding area in a steel strip surface image using the grinding area detection model; and periodically detecting the grinding area using the grinding area detection model to obtain a grinding area detection result.

2. The steel strip surface grinding image detection system as described in claim 1, wherein the grinding area detection model is established using an artificial neural network (ANN) method.

3. The steel strip surface grinding image detection system as described in claim 1, wherein the grinding area detection model includes at least one of a weighted bidirectional feature pyramid network (BiFPN) and a region proposal network (RPN).

4. The steel strip surface grinding image detection system as described in claim 1, wherein the processor is further configured to display the detection result of the grinding area on a user interface.

5. The steel strip surface grinding image detection system as described in claim 4, wherein the processor is further configured to query the grinding area detection results based on a date, a production line type, and a steel coil number input by a user, and display the queried grinding area detection results on the user interface.

6. The steel strip surface grinding image detection system as described in claim 5, wherein the grinding area detection result includes at least one of a grinding area location, a grinding area bounding box, and a detection accuracy.

7. The steel strip surface grinding image detection system as claimed in claim 6, wherein the processor is further configured to set a filtering threshold based on the detection accuracy, and to use the filtering threshold to determine the at least one grinding image to be displayed on the user interface.

8. The steel strip surface grinding image detection system as claimed in claim 1, wherein the processor is further configured to store the grinding area detection result output by the grinding area detection model into a detection result database, wherein the detection result database includes at least one form field, and wherein the at least one form field includes at least one of a read image path field, a save image path field, a recognition quantity field, and a recognition result confidence value field.

9. A method for detecting a steel strip surface grinding image, executed by a processor, comprising: accessing a grinding image database and obtaining at least one grinding image, then establishing a grinding region detection model based on the at least one grinding image; using the grinding region detection model to locate a grinding region in a steel strip surface image; and using the grinding region detection model to periodically detect the grinding region to obtain a grinding region detection result.

Images