In order to realize soil compactness measurement,a laser image measurement system is established and the non-destructive measurement method of soil compactness is investigated.Firstly,the laser image of soil is smoothed by using the 4-neighborhood average filter algorithm.Secondly,the laser spot of the laser image is extracted by using Canny algorithm.Then the water content,laser spot radius,the absorption coefficient and scattering coefficient are selected as the classifier input feature values.Finally,the soil compactness is predicted by BP neural network.The experimental results show that BP neural network reaches the required error after 11 loops.Compared with the measurement values by using the round knife method,the average absolute error and relative error of compactness are less than 2%.Therefore,the measurement accuracy of the system can meet the requirements.