Image-based real-time displacement measurement system

A precise displacement measurement system using a non-contact image-based optical sensor and a hardware-based artificial neural network was developed. Field tests, with the sensors mounted on a tractor-drawn toolbar, were conducted to compare the performance of the image-based displacement system with traditional ground-wheel driven encoder-based and radar-based displacement systems at two different travel speeds (0.45 and 1.36 m s−1). The precision of the three methods, expressed as the standard deviation of spot spray activation at 127 mm intervals, was 7.3 mm, 5.9 mm, and 17.3 mm on a rough soil surface (50 mm to 125 mm sized soil clods) at a 0.45 m s−1 travel velocity for the image-based, wheel-based, and radar-based displacement sensors respectively. Compared to the radar-based sensor, the image-based sensor had better accuracy on a rough soil surface than the wheel-based sensor, better accuracy at both speeds, and better precision at the low speed. In addition to being non-contact, the main advantage of the image-based sensor over the wheel-based sensor was that it did not require site specific re-calibration to maintain accuracy.