Silicon imagers with integrated motion-detection circuitry have been develo
ped and tested for the past 15 years. Many previous circuits estimate motio
n by identifying and tracking spatial or temporal features. These approache
s are prone to failure at low SNR conditions, where feature detection becom
es unreliable. An alternate approach to motion detection is an intensity-ba
sed spatiotemporal correlation algorithm, such as the one proposed by Hasse
nstein and Reichardt in 1956 to explain aspects of insect vision. We implem
ented a Reichardt motion sensor with integrated photodetectors in a standar
d CMOS process. Our circuit operates at sub-microwatt power levels, the low
est reported for any motion sensor. We measure the effects of device mismat
ch on these parallel, analog circuits to show they are suitable for constru
cting 2-D VLSI arrays. Traditional correlation-based sensors suffer from st
rong contrast dependence. We introduce a circuit architecture that lessens
this dependence. We also demonstrate robust performance of our sensor to co
mplex stimuli in the presence of spatial and temporal noise.