Detecting object region and working state of aerator based on computer vision and machine learning
share
Aerator plays an important role in the regulation of dissolved oxygen in aquaculture. The development of computer vision technology provides an opportunity for realizing intelligent monitoring of aerator. Surveillance cameras have been widely used in aquaculture. Therefore, it is of great application value to detect the working state of the aerator with the existing surveillance cameras. In this paper, a method of object region detection and working state detection for aerator is presented. In the object region detection module, this paper proposes a method to detect the candidate region and then determine the object region, which combines the background modeling, the optical flow method and the maximum inter-class interval method. In the work state detection module, this paper proposes a novel method called reference frame Kanade-Lucas-Tomasi (RF-KLT) algorithm, and constructs a classification procedure for the unlabeled time series data. The results of this study show that the accuracy of detecting object region and working state of aerator in the complex background is 100 detection speed is 77-333 frames per second (FPS) according to the different types of surveillance camera. Compared with various foreground detection algorithms and machine learning algorithms, these methods can realize on-line, real-time and high-accuracy detection of the object region and working state of aerator.
READ FULL TEXT