HUBEI AGRICULTURAL SCIENCES ›› 2022, Vol. 61 ›› Issue (20): 195-202.doi: 10.14088/j.cnki.issn0439-8114.2022.20.037

• Information Engineering • Previous Articles     Next Articles

Video shot detection algorithm of fruit pest knowledge based on multi feature fusion

LIU Li2, FENG Hong-cai1b, HUANG Qing1a   

  1. 1a. School of Mathematics and Computer Science; 1b. Network and Information Center, Wuhan Polytechnic University, Wuhan 430023, China;
    2. Department of Information Technology, Wuhan Dongxihu Vocational and Technical School, Wuhan 430023, Chin
  • Received:2021-12-08 Online:2022-10-25 Published:2022-11-23

Abstract: In order to improve the accuracy of shot detection in agricultural knowledge video, taking fruit pest knowledge video as an example, an adaptive double threshold fast shot detection method based on multi-feature fusion was used. The method was divided into two stages: initial inspection and reinspection. In the initial inspection stage, the improved block was used to extract the color features according to the characteristics of human vision, and the candidate frame sequence of the shot boundary was selected combined with the adaptive skip method; in the inspection process, the multi feature fusion method was used to further refine the video shot boundary obtained from the initial inspection. The results showed that this method could effectively realize the rapid segmentation of fruit diseases and insect pests knowledge video shot, with a recall rate of more than 93.5% and a precision rate of more than 93.4%. It could efficiently and accurately find the shot of fruit diseases and insect pests, and improve the utilization rate of farmers'agricultural knowledge video.

Key words: fruit diseases and insect pests knowledge video, shot segmentation, lens similarity, multi feature fusion, double threshold

CLC Number: