Development of a device for estimating the optimal artificial insemination time of individually stalled sows using image processing.

摘要

This study was conducted to develop an artificial insemination time estimator (OAITE) for individually stalled sows using image processing and evaluate the performance of the OAITE through field test. The OAITE was consisted of a computer, multiplexer, three CCD cameras and three LED lamps (950 nm wavelength). The computer programme used for the OAITE to quantify the lying and non-lying (sitting and standing) rates of sows in stalls was written in LabWindows/CVI. For the purpose of establishing references that would help estimate the optimum artificial insemination (AI) time for sows, the lying rate of 50 Berkshire x Hampshire crossbred sows (parity: 2 to 7) was observed and recorded. The observation was made on the second day after the sows were moved into the stalls when they were artificially inseminated. Results revealed that the lying rates on the day of oestrus and the other days were not significantly different when based on the following time bands: 00, 08, 09, 16, and 17 (p>0.1). Thus, only the time bands other than these time bands were used to establish the references for determining the onset of the oestrus. Based on the lying rates observed and the references established by the procedures above, the study assigned 0 to the lying rate of the non-oestrus time band, 0.5 to the lying rate between the non-oestrus and oestrus time bands and 1 to the lying rate of the oestrus time band. The authors assumed that if the OAITE produced 0.5 or above >four times in a row and if the results included 1 at least once, the oestrus would have started. In addition, it was assumed that the optimum AI time for sows was between 26 and 34 h after the beginning of oestrus. The results of sows' AI of the OAITE group (n=40 sows; AI=1 time) showed that the pregnancy rate was 92.5%, which was the same rate as the control group (n=40 sows; AI=2 times), and that the litter size did not differ between the control and the OAITE group. It is suggested that the OAITE is effective and cost efficient to estimate the optimum AI time of individually stalled sows.