Artificial intelligence (AI) has become the new technique for solving the most complicated challenges related to big data, image recognition, object detection, and prediction issues due to the fourth industrial revolution. This article introduces the development and implementation of a zinc-plated component recognition system within a manufacturing process using deep learning (DL) techniques. This paper aims to train and evaluate different DL algorithms to recognize five zinc-plated components and one assembly tray under different ambient light conditions and finishing. The proposed method begins with creating a custom dataset of 5.5K images, with six classes that match the assembly components. However, before the training stage starts, a subset between 5% and 10% is created. Eight deep neural networks have been benchmarked: Swing Transformer, Convnextv2, Xception41, Data-efficient image Transformer, Inception_v4, ViT, Resnet50 and Efficientnet_b0. The training stage is performed with the k-fold method using k = 6 and a progressive data size between different runs with an average of data utilized during training of 8%. In this study, convnextv2_tiny exhibits the highest F1-Score with 99.177%, followed by swin_tiny_patch4_window7_224 with 99.175%, while resnet50 achieved the lowest F1-Score at 95.886%. Therefore, the new generation of convolutional architectures is still better for classifying images. Thus, these algorithms can be adopted as good classifier models for zinc-plated component classification in the industrial manufacturing environment.