| As the world’s superpower of pig’s farming, large scale and industrialization pig’s farming has developed rapidly in China, but the disease is a major problem has restricted the industry of high efficiency and high yield. Panting pathological movement is a common symptoms when pig suffering from respiratory diseases, and can be used as morbid warning basis. So intelligent monitoring to panting behavior based on computer vision technology has important research value and application prospect. According the monomer ridge abdominal contour significant changes when panting behavior happens in the video and the RACID method has shortcoming, this paper studies the building pant behavior method of one dimensional wave graph model. The specific research work are as follows:(1) the mean curvature radius get ridge of contour extraction method of ACR is studied, To locate ridge contour fluctuations on the most significant part, positioning of the ridge of contour algorithm effectively based on the envelope is studied;(2) a method of maximum inscribed circle diameter MICD in pig’s outline is studied, In order to reduce the leg area’s calculation complexity when extracting maximum inscribed circle in pig’s outline, a algorithm based on outline Harris corner to remove the legs is studied;(3)a method of using of two video structure characteristic parameters(ACRD and MICDD) building breath behavior’s waveform graph model is introduced, In order to quantitatively describe pigs pant degree of abnormal breathing, Pig respiratory frequency calculation method is proposed;The method can be used for continuous monitoring, quantify the pant behavior of pig farms, By saving a rhythmic breathing waveform graph information, the method can automatically monitor early warning panting behavior suspected sick pigs. The method in this paper can open up a new diagnosis approach of animal body condition and improve the level of intelligent pig breeding technology on the basis of applicable significance. |
PDF Full Text Request