Chromatic confocal sensor is widely used in various precision measurement fields due to its advantages of high measurement accuracy, fast response speed and good stability. Different from the traditional fiber-coupled structure, we propose an integrated compact chromatic confocal sensing system that allows to overcome the device integrating constraints met in industrial environments. Aiming at the distortion of the peak waveform caused by the inconsistent spectral response of the system and to accurately extract the peak wavelength, a spectral characteristic compensation algorithm and a peak wavelength extraction method based on Gaussian curve fitting are proposed. Based on these, a segmented curve calibration algorithm is put forward to achieve accurate mapping between peak wavelength and position. For the thickness measurement of transparent objects, a simple thickness measurement model and its calibration procedure are proposed, which do not need to obtain prior parameters such as incident angle or refractive index. Finally, the performance of the proposed sensing system is tested by displacement measurement experiment and thickness measurement experiment. The experimental results show that the RMSE of displacement measurement is less than 0.1μm, and the RMSE of thickness measurement is less than 1μm, which verifies the effectiveness and feasibility of the proposed sensing system.