diff --git a/modules/gpu/src/stereobp.cpp b/modules/gpu/src/stereobp.cpp index 3b827a3129..dde69d6d44 100644 --- a/modules/gpu/src/stereobp.cpp +++ b/modules/gpu/src/stereobp.cpp @@ -158,7 +158,7 @@ namespace init(stream); - datas[0].create(rows * rthis.ndisp, cols, rthis.msg_type); + createContinuous(rows * rthis.ndisp, cols, rthis.msg_type, datas[0]); comp_data_callers[rthis.msg_type == CV_32F][left.channels()](left, right, datas[0], StreamAccessor::getStream(stream)); @@ -187,10 +187,10 @@ namespace private: void init(Stream& stream) { - u.create(rows * rthis.ndisp, cols, rthis.msg_type); - d.create(rows * rthis.ndisp, cols, rthis.msg_type); - l.create(rows * rthis.ndisp, cols, rthis.msg_type); - r.create(rows * rthis.ndisp, cols, rthis.msg_type); + createContinuous(rows * rthis.ndisp, cols, rthis.msg_type, u); + createContinuous(rows * rthis.ndisp, cols, rthis.msg_type, d); + createContinuous(rows * rthis.ndisp, cols, rthis.msg_type, l); + createContinuous(rows * rthis.ndisp, cols, rthis.msg_type, r); if (rthis.levels & 1) { @@ -216,10 +216,10 @@ namespace int less_rows = rows / 2; int less_cols = cols / 2; - u2.create(less_rows * rthis.ndisp, less_cols, rthis.msg_type); - d2.create(less_rows * rthis.ndisp, less_cols, rthis.msg_type); - l2.create(less_rows * rthis.ndisp, less_cols, rthis.msg_type); - r2.create(less_rows * rthis.ndisp, less_cols, rthis.msg_type); + createContinuous(less_rows * rthis.ndisp, less_cols, rthis.msg_type, u2); + createContinuous(less_rows * rthis.ndisp, less_cols, rthis.msg_type, d2); + createContinuous(less_rows * rthis.ndisp, less_cols, rthis.msg_type, l2); + createContinuous(less_rows * rthis.ndisp, less_cols, rthis.msg_type, r2); if ((rthis.levels & 1) == 0) { @@ -286,7 +286,7 @@ namespace cols_all[i] = cols_all[i-1] / 2; rows_all[i] = rows_all[i-1] / 2; - datas[i].create(rows_all[i] * rthis.ndisp, cols_all[i], rthis.msg_type); + createContinuous(rows_all[i] * rthis.ndisp, cols_all[i], rthis.msg_type, datas[i]); data_step_down_callers[funcIdx](cols_all[i], rows_all[i], rows_all[i-1], datas[i-1], datas[i], cudaStream); }