| Real-time monitoring of the weight of fattening pig is an important criterion for evaluating the growth and health of fattening pig. Now the measurement of body weight depends mainly on the loadometer weighing. That is not only stress response consequences for pigs, but also labor-intensive and time-consuming for human. Therefore, it is important significant to study the weight of pigs without contact. This article is based on the analysis of domestic and foreign research status, measured the weight of fattening pigs without contact by using the computer vision technology.According to the characteristic of the pig’s images in this paper, the image is preprocessed, to reduce the influence of light and noise. Then, a new feature model is established by using the color feature, get a new sample image.It can make it easier to image threshold segmentation and morphological image processing, compared the image processing effect of three different methods that includes in histogram, iterative and OTSU method. Select the best segmentation method for this passage, get the binary image of pig that has better shape, and provide convenience for the feature extraction of pig. Design the estimation system of pig’s projection area based on the GUI, it can realize the human-computer interaction, make our life more convenient and facilitate the measurement of pig’s size.Count the data of pig’s size and weight. By the comparison of the fitting correlation of the least square method, stepwise regression method, volume model method and MLP neural network model, we find that the correlation of the MLP neural network model is the best. The correlation coefficient is 0.99, and the average relative error is 1.18%. It can assure the precision of evaluations very well, and provide new theoretical basis for measuring the weight of fattening pig. |
