Method for increasing the information value of video data based on the removal of redundant frames and entropy estimation

The use of modern video surveillance systems is associated with solving the tasks of monitoring the activities of personnel and compliance with the technological process based on the analysis and processing of large amounts of video data. This leads to an increase in the cost of information storage, the cost of staff time resources to search for key events over long time periods. The problem of increasing the information value of stored data from video surveillance cameras based on frame filtering and entropy estimation is considered. The implementation of algorithms for processing and compressing information aimed at reducing the volume of stored video data is proposed. The use of this implementation contributes to increasing the overall information value, the efficiency of video surveillance systems by optimizing the volume of stored information and increasing the ratio of useful information. To increase the informational value of video data, a method is proposed that includes the use of modern video compression technologies, a frame filtering algorithm, and an evaluation of the processed video by the Shannon entropy metric. The analysis and comparison of existing video data compression algorithms are performed. An experiment was carried out, as a result of which the correlation between high entropy values and the information value of the frame was proved, the frame filtering algorithm was successfully tested, which allowed to increase entropy by 5.4 times and reduce the duration of the video by 8 times. The use of video data compression methods and efficient codecs, for example, H.265/HEVC, reduced the file size by 14.57 times compared to the original one. The approbation of the proposed method is considered when solving problems of filtering, transmitting, and storing of video data to increase the information value of video data, the productivity of the analysis and information retrieval processes by reducing redundant, useless data fragments. The advantage of the presented method is to remove redundant frames based on motion analysis and entropy estimation of video data, a combination of various approaches to reduce the volume of transmitted and stored information. The application of the method will increase the efficiency of data storage in various video surveillance systems (for logistics centers, warehouse complexes, retail premises).