Endmember (EM) selection is an important prerequisite task for mixedspectral analysis of hyperspectral imagery. In all kinds of EM selection methods,N-FINDR has been a popular one for its full automation and efficient performance.Unfortunately, the implementation of the algorithm needs dimensional reduction inoriginal data, and the algorithm includes innumerable volume calculation. Thisleads to a low speed of the algorithm and so becomes a limitation to itsapplications. In the present paper, an improved N-FINDR algorithm was proposedbased on linear least square support vector machines (LLSSVM), which is free ofdimensional reduction and makes use of distance measure instead of volumeevaluation to speed up the algorithm. Additionally, it was also proposed to endowthe algorithm with robustness by controlling outliers. Experiments show that thecomputational load for EM selection using the improved N-FINDR algorithm based onLLSSVM was decreased greatly, and the selection effectiveness and the speed ofthe proposed algorithm were further improved by outlier removal and the pixelpre-sorting method respectively.support vector machines(LLSSVM); N-FINDR algorithm水下智能机器人技术国防科技重点实验室项目资助