Abstract: In order to solve the "mosaic" problem in plateau histogram enhancement algorithms,a novel method was proposed based on statistics,reuse of accumulated space and lossless compression.After theoretical analysis and simulation,the new algorithm reduces memory requirement more than 85%,but the time complexity keeps the same level as the original plateau histogram equalization.As a result,the processing space of plateau histogram equalization is greatly compressed by this improved algorithm,the difficulty of algorithm realization caused by the shortage of running memory of operating chips in infrared real time imaging system is solved and the requirements of practical application are achieved.