With the popularity of the Internet and the development of digital technology, the amount of image and video data content in human resources management in colleges is also increasing. It is very important to achieve effective management and retrieval of video image data. To this end, this paper takes clustering algorithm as the research object, discusses the image processing of clustering algorithm and its application in human resources management of colleges and universities, in order to provide reference for my country's colleges and universities to realize the optimal management of human resources. Based on the content of digital images, this paper introduces three traditional digital image calculation methods and calculates their calculation process, including SIFT algorithm, VALAD algorithm and PCA algorithm. On this basis, the traditional algorithm is improved, and the improved Algorithm; and then take a college as the specific research object, through the use of K-means algorithm and improved K-means algorithm to carry out cluster analysis of the school's professional and technical personnel database, and formulate new Standard, that is, the working hours of the 5th and 6th levels of the sub-high school must be 402 and 378 months respectively, the appointment of the sub-high school must be 165 and 98 months, the total class hours are 1026 and 789 hours, and the scientific research scores are 1402.4 and 785.3 points.. The research results show that the clustering algorithm has a significant role and significance for the optimal management of human resources in universities and other enterprises.