The dynamic range and exposure time of the camera are the main limitation for the traditional optical projection tomography to achieve the intact and detailed three-dimensional information of samples with complicated spatial structure. The normalized dynamic range-transform of three-dimensional imaging acquisition and process are used to solve the existing conundrum of traditional three-dimensional imaging based on the optical projection tomography technology, by means of setting the Lambert source and carrying on the new method of normalized dynamic range-transform. Setting multiple exposures to get the image of sample, normalizing the tested response curve of the camera for solving the non-linear distortion and false appearance existing in the traditional high dynamic range, using the image fusion theory to improve the image processing program, then reconstructing the sample by the reverse projection algorithm, and get the three-dimensional structure of the sample finally. From the analysis of imaging result, the normalized dynamic range-transform of three-dimensional imaging acquires more information of samples with complicated spatial structure, more details of different areas are able to be observed through this way.