Abstract: Abstract: The quality of seedbed after seeding such as soil surface leveling, soil-breaking rate and seed furrow preparing and so on, are the main factors affecting the seedling rate of rapeseed. A good seedbed is a necessary condition for uniform seedling emergence, increasing the rate of subsequent seedlings and promoting the homogeneity of plants after sowing. In view of the fact that the measurement and data processing methods of rapeseed seedbed surface roughness under mechanized direct seeding cannot fully reflect the real soil surface condition, a method of measuring and quantifying roughness surface area by using laser radar scanning technology is proposed in this paper. The main components of the measuring device are the lidar, linear guide, stepper motor, portable computer and other components. The device can quickly acquire soil surface data in three dimensions through the whole working width of direct seeding machine. The surface roughness data of rapeseed combined direct seeder and no-tillage direct seeder after operation were used to detect the effect of two kinds of seedbed operation on seedbed surface. The characteristics of rapeseed direct seeding process were analyzed, three-dimensional digital modeling and detrend processing of ground data after two kinds of direct seeders were carried out by using MATLAB. The least square method was utilized to eliminate the inclined tendency of seedbed surface, the fourth order polynomial was applied to remove the slope trend in the width direction of the soil surface of rapeseed combined direct seeder. The root mean square (RMS) height and the correlation length were involved as evaluating parameters to analyze the data. Statistical results of roughness parameters for different spatial sampling intervals and different angles cross-section data showed that the average RMS height and RMS height mean error of 170 mm sampling interval were all higher than that of 5 mm sampling interval, the average RMS height of the 170 mm sampling interval was 2.2 mm higher than that of the 5 mm sampling interval, the corresponding correlation length was around 1.8 times the latter; and the RMS height and the correlation length error of the 170 mm sampling interval were higher than that of the 5 mm sampling interval. The maximum difference of RMS height and correlation length of surface elevation data were 7.69 and 25.14 mm respectively in the three directions of 0, 45 and 90 degrees in the forward direction of vertical unit, and there was obvious anisotropy on the surface of seedbed with structural characteristics such as seed trench. Statistical results of local roughness quantification using sliding sampling windows in different size regions showed that when the window width is 0.5 times the compartment width, the standard deviation of RMS height can be stabilized within 0.27 mm if the length of the window is not less than 1.2 m, and the influence of local surface tilt on the roughness calculation can be eliminated by the trend treatment of the tilt of each sampling window. The results can provide reference for the measurement and quantitative analysis of surface roughness in rape direct seeding operation.