通过设置class_doit 的参数 std_mult=3.0 即可。
完整代码如下：
参考引用：http://hi.baidu.com/dyqwrp/item/c8318cea9cec732c5a7cfbab

pro Parallelpiped_methodCOMPILE_OPT idl2;恢复ENVI_SAVE文件ENVI,/restore_base_save_files;开始批处理模式ENVI_BATCH_INIT,log_file='batch.log';编写批处理过程;选择文件inputfile=   'E:\海二相关\掩膜\jpproject.img'CATCH, Error_statuserrorshow = 'Sorry to see the error,'+ $' please send the error Information to "myyouthlife@163.com"'IF Error_status NE 0 THEN BEGINtmp = DIALOG_MESSAGE(errorshow+STRING(13b)+$!ERROR_STATE.MSG,/error,title = '错误提示!')returnENDIF;输入数据预处理ENVI_OPEN_FILE, inputfile, r_fid=fidIF (fid EQ -1) THEN BEGINRETURNENDIF;获取输入文件信息ENVI_FILE_QUERY, fid, dims=dims, nb=nbpos  = LINDGEN(nb)
;     输出分类图像名称outputfile='d:\result.img'out_name = outputfile;-平行六面体分类  0ENVI_RESTORE_ROIS, 'E:\海二相关\掩膜\11.roi'
;      获取感兴趣区文件信息roi_ids = ENVI_GET_ROI_IDS(fid=fid, $roi_colors=roi_colors, roi_names=class_names)class_names = ['Unclassified', class_names]num_classes = N_ELEMENTS(roi_ids); Set the unclassified class to black and use roi colorslookup = BYTARR(3,num_classes+1)lookup[0,1] = roi_colors; 计算类ROI的基本统计信息mean = FLTARR(N_ELEMENTS(pos), num_classes)stdv = FLTARR(N_ELEMENTS(pos), num_classes)cov = FLTARR(N_ELEMENTS(pos),N_ELEMENTS(pos),num_classes)FOR j=0, num_classes-1 DO BEGIN;roi_dims=[ENVI_GET_ROI_DIMS_PTR(roi_ids[j]),0,0,0,0]ENVI_DOIT, 'envi_stats_doit', fid=fid, pos=pos, $dims=roi_dims, comp_flag=4, mean=c_mean, $stdv=c_stdv, cov=c_covMEAN[0,j] = c_meanstdv[0,j] = c_stdvcov[0,0,j] = c_covENDFOR;设定阈值thresh=REPLICATE(0.1,num_classes)out_bname = 'parallelepiped'
;      分类ENVI_DOIT, 'class_doit', fid=fid, pos=pos, dims=dims, $out_bname=out_bname, out_name=out_name, method=0, $mean=mean, stdv=stdv, std_mult=3.0, $lookup=lookup, class_names=class_names, $r_fid=r_fid,in_memory=0; ;      退出批处理模式;envi_batch_exit
end