Research on Segmentation Algorithm of Workpiece Character Image Based on FCN-ELM

In an industrial scene, in order to deal with the character segmentation problem on each workpiece, multiple complex steps such as image preprocessing are required, and then image segmentation processing of the workpiece character is performed. Aiming at the complex preprocessing process in the workpiece character recognition process, a workpiece character image segmentation algorithm based on the combination of FCN and ELM is proposed, that is, the convolution layer is mainly responsible for the feature extraction function in the FCN, and the difference of the input image is extracted Deep image features, remove the last fully connected layer of the convolutional neural network, and then use ELM instead of the fully connected layer, as a classifier to quickly classify image pixels, improve the classification accuracy, and then connect the convolutional layer to the extreme learning machine. The classified image pixels are used as the input of the convolutional layer to perform an up-sampling operation, and the generated feature thumbnails need to be restored to the dimensions of the original image. The experimental results show that the improved algorithm improves the accuracy compared with the original FCN algorithm.