To inspect weight and growth of pigs in real time，an web-based inspecting system based on binocular stereovision technology was developed. Hardware and software related to this inspecting system were established in laboratory condition. The calibration algorithm with non-linear camera model and coordinate inspecting arithmetic based on least square methods was developed. Calibration mode about different numbers, parts and rotating angle of the calibration board images was collected and the inspecting results were analyzed. Results showed that more than 19 calibration board images should be used to get steady system parameters. Collecting the calibration images in all parts of inspecting zone could improve the inspecting precision. There was no obvious difference in inspecting precision at different rotating angle, but a large rotating angle was not conducive to extract all coordinates of the calibration points.