Research On The Machine Vision Methods For Roughness Detection Based On Color Information

The surface roughness of a workpiece plays an important role in mechanical properties,including its fitting,wear and anti-corrosion and its anti-fatigue,contact stiffness,vibration and noise,etc.,such as closely related to the service life and the reliability of mechanical product.The existing machine-vision surface roughness measurement technique extracts relevant evaluation indices from grayscale images without using the strong sensitivity of color information or considering the subjective evaluation of the human visual system.In addition,most of these measurements use a micro-vision imaging method to measure a small area and cannot make an overall assessment of the workpiece's surface.Additionally,the micro-vision measurements method requires a microscopic device,causing small test area,inconvenient operation and low efficiency,and cannot work on-line.In addition,the laser optical systems are very expensive,the measurement equipment is complex in structure,and optical path adjustment is rather difficult.To address these problems,The detection method of surface roughness based on ordinary light source,non micro-vision and color information was studied in this paper.Based on the difference in sharpness of virtual images of color blocks that are formed on the workpiece's surfaces with different roughness,an algorithm for no reference evaluating the sharpness of color images that is based on the RGB color space was used to develop a correlation model between the sharpness and the surface roughness,and the feasibility of the detection method based on the sharpness of color image is demonstrated.The results show that the sharpness is strongly correlated with the surface roughness,and relative to the detection method for surface roughness using gray image,the proposed method is convenient,highly accurate and conform to the subjective evaluation of the human vision system.Based on the good performance for sharpness evaluation index of color image to roughness characterization,and starting from the mechanism research of indices,the average color difference index which physical meaning is clear was proposed.A comparison test is conducted on a set of test samples before and after surface contamination using the color index and gray-level algebraic averaging,the square of the main component of the Fourier transform in the frequency domain,and the entropy.A strong correlation between the color index and the surface roughness is established;this correlation is not only higher than that of other indices but also present despite contamination and very robust.Verification using a regression model based on a support vector machine proves that the proposed method not only has a simple apparatus and makes measurement easy but also provides high precision and is suitable over a wide measurement range.The impact of the red and green color blocks,the lighting,and the direction of the surface texture on the correlation between the color index and the roughness are also assessed and discussed in this paper.The results indicate the relationship between the lighting and color difference index is linear,and the color block's design can alter the impact of texture on roughness measurement.Aiming at the universal applicability for the average color difference index to the surface roughness processed by the different processing technology,A discussion about different texture roughness parameters using the color difference index and five typical spectrum index was conducted.The research results show that the average color difference index has a stronger correlation with surface roughness and a better robustness to the incident angle than spectrum index,and the reason was analyzed why the correlation of index and milling sample is weaker than grinding sample from the mechanism.Based on the customized red-green checkerboard pattern light source,and the discussion of the sum of squared residuals through two texture direction of workpiece,studies suggest when the camera angle relative to the workpiece's surface normal is small,the coincidence degree of relation curve between two direction is high,which provides the implementation basis for ignoring texture direction of sample surface to automation detection based on machine vision.In addition,to address the problem whether the correlations are consistent between average color difference index and different material roughness under the same processing technology,paper uses the customized red-green light source to study five kinds of materials in different angle of incidence.The results show that correlation performance between different material roughness and color difference index in different incident angle is stable.But the difference of the mathematical relationship model between index and roughness proves that materials have an effect on color information detection roughness in mathematical nature.Ambient light noise research through different materials indicates that color difference index is robust,not only to the incident angle but also to ambient light noise.Through the expand research on color image feature index,and based on the multi-dimensional information of color image and the mixing structure of red and green color,the color distribution statistical matrix was built where the aliasing area index and the energy index in pure color area were extracted to characterize roughness.The research results show that the aliasing area index can be used in high precision grinding roughness detection,and the energy index in pure color area shows very good universality as same as the color difference index.In addition,the performance evaluation system of roughness characteristics index was established from the perspective of instrument development,which synthetically evaluated indices in terms of accuracy,monotonicity,stability and efficiency of performance.Relative to the index of gray image,the color information index has a better performance in various aspects,and the machine vision and automation method for roughness detection based on color information and ordinary light is laid a solid foundation of theory and experiment.