In order to improve the accuracy of end-to-end text detection algorithm, a steel coil inkjet detection algorithm based on high generalization ABCNet is proposed. In the mainstream ABCNet text detection algorithm, the Backbone module is optimized and the high generalization ability is integrated to improve the detection accuracy. Experiments show that the algorithm can effectively improve the detection accuracy of steel coil inkjet coding, and compared with EAST, TextSnake, DBNet + +, PAN-640 and other models in terms of accuracy Precision and F-Measure. The results show that the algorithm has obvious advantages. On the steel coil detection dataset, Precision and F-Measure achieved good results, Precision was 99.7%, and F-Measure was 99.2%.