fixed bug in gpu filter engine (incorrect buffer type) and in vector's saturate_cast.

changed buffer type in linear filters to float.
added support of 1 channel image to linear filters.
added support of BORDER_REFLECT101, BORDER_REPLICATE and BORDER_CONSTANT border type to gpu linear filters.
minor fix in tests.
update comments in gpu.hpp.

modules/gpu/src/imgproc_gpu.cpp

@@ -972,6 +972,12 @@ bool cv::gpu::tryConvertToGpuBorderType(int cpuBorderType, int& gpuBorderType)
         gpuBorderType = cv::gpu::BORDER_REPLICATE_GPU;
         return true;
     }
+    
+    if (cpuBorderType == cv::BORDER_CONSTANT)
+    {
+        gpuBorderType = cv::gpu::BORDER_CONSTANT_GPU;
+        return true;
+    }
 
     return false;
 }