传统体素重建方法的弊端是用测量数据重建庞大的光学参量。庞大的数据量一方面 加大重建过程中的计算量，另一方面由于测量数据易于受噪音的影响造成结果的不稳定性。因此， 本文提出一种基于区域的重建方法，该方法通过找邻近节点确定目标区域。第一步进行体素重建， 确定目标个数及其中心点。第二步重建中心点所在的区域。实验中，给出了重建模型吸收系数的重 建结果，并引入四个数值指标进一步衡量重建图像质量。实验结果表明: 该方法可以有效降低重建 系统规模，克服重建病态性，提高重建图像的精确性和对比度。

A common difficulty for the traditional methods of diffuse optical tomographic re-construction is that only a small amount of measurements can be used to recover the image com-prised of a large number of pixels, which not only leads to expensive computational cost but also results in an unstable solution prone to be affected by the noise in the measurement data. In this paper, a region-based method for reducing the unknowns is proposed, where the target areas are determined by searching for the nearest neighbor nodes. The first is to determine the number of targets and their centers, then to reconstruct the regions from the initial estimated centers. Re-constructed results corresponding to the absorption coefficients are given. Four numerical criteria are imported to evaluate the quality of reconstructed images. Experimental results show that the reconstructing system scale can be decreased dramatically and problem of ill-posedness fixed. At the same time , the accuracy and contrast of reconstructed images are improved.