Multi-modal semantic segmentation method, system and device in industrial scene and storage medium

The invention relates to the field of semantic segmentation technologies, in particular to a multi-modal semantic segmentation method, system and device in an industrial scene and a storage medium. The multi-modal semantic segmentation method in the industrial scene comprises the steps of obtaining an RGB detection image containing a to-be-recognized object and text data of the to-be-recognized object; inputting the RGB detection image and the text data into a semantic segmentation model, respectively extracting image features of the RGB detection image and text features of the text data through the semantic segmentation model, and aligning and fusing the image features and the text features to obtain processed semantic features, and judging the probability that the RGB detection image is a target object based on the semantic features as output to obtain a recognition result of the RGB detection image, so that the semantic segmentation effect of the target object to be recognized in a complex industrial scene