Starting from the communication characteristics of piano music culture in colleges and universities, this paper proposes a piano music signal detection technology that includes two parts: note onset and fundamental frequency detection. Based on the characteristics of the harmonic structure of the music signal, the degree of harmonic matching between the signal to be recognized and the standard note signal is calculated, and the task of multi-fundamental frequency estimation is accomplished by using the timbre characteristics and harmonic structure. The detection technology of piano music signals is applied to the recognition of teaching short notes, playing music and playing evaluation to get a new model of piano music teaching in colleges and universities. The results show that the algorithm in this paper can recognize piano music correctly by more than 91%, and the method has no obvious decrease in recognition rate in the process of gradually increasing noise, which is conducive to improving the effect of piano music teaching in colleges and universities.