generic-library/opencv
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

cv::cvtColor (RGB2HLS CV_8U)

Browse Source
This commit is contained in:
Ilya Lavrenov 2014-10-07 15:33:09 +00:00
parent 8d367d4b4d
commit c173d9fa18
1 changed files with 63 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

66
modules/imgproc/src/color.cpp
View File

@ -3208,7 +3208,14 @@ struct RGB2HLS_b
typedef uchar channel_type;
RGB2HLS_b(int _srccn, int _blueIdx, int _hrange)
: srccn(_srccn), cvt(3, _blueIdx, (float)_hrange) {}
: srccn(_srccn), cvt(3, _blueIdx, (float)_hrange)
{
#if CV_NEON
v_scale_inv = vdupq_n_f32(1.f/255.f);
v_scale = vdupq_n_f32(255.f);
v_alpha = vdup_n_u8(ColorChannel<uchar>::max());
#endif
}
void operator()(const uchar* src, uchar* dst, int n) const
{
@ -3218,8 +3225,41 @@ struct RGB2HLS_b
for( i = 0; i < n; i += BLOCK_SIZE, dst += BLOCK_SIZE*3 )
{
int dn = std::min(n - i, (int)BLOCK_SIZE);
j = 0;
for( j = 0; j < dn*3; j += 3, src += scn )
#if CV_NEON
for ( ; j <= (dn - 8) * 3; j += 24, src += 8 * scn)
{
uint16x8_t v_t0, v_t1, v_t2;
if (scn == 3)
{
uint8x8x3_t v_src = vld3_u8(src);
v_t0 = vmovl_u8(v_src.val[0]);
v_t1 = vmovl_u8(v_src.val[1]);
v_t2 = vmovl_u8(v_src.val[2]);
}
else
{
uint8x8x4_t v_src = vld4_u8(src);
v_t0 = vmovl_u8(v_src.val[0]);
v_t1 = vmovl_u8(v_src.val[1]);
v_t2 = vmovl_u8(v_src.val[2]);
}
float32x4x3_t v_dst;
v_dst.val[0] = vmulq_f32(vcvtq_f32_u32(vmovl_u16(vget_low_u16(v_t0))), v_scale_inv);
v_dst.val[1] = vmulq_f32(vcvtq_f32_u32(vmovl_u16(vget_low_u16(v_t1))), v_scale_inv);
v_dst.val[2] = vmulq_f32(vcvtq_f32_u32(vmovl_u16(vget_low_u16(v_t2))), v_scale_inv);
vst3q_f32(buf + j, v_dst);
v_dst.val[0] = vmulq_f32(vcvtq_f32_u32(vmovl_u16(vget_high_u16(v_t0))), v_scale_inv);
v_dst.val[1] = vmulq_f32(vcvtq_f32_u32(vmovl_u16(vget_high_u16(v_t1))), v_scale_inv);
v_dst.val[2] = vmulq_f32(vcvtq_f32_u32(vmovl_u16(vget_high_u16(v_t2))), v_scale_inv);
vst3q_f32(buf + j + 12, v_dst);
}
#endif
for( ; j < dn*3; j += 3, src += scn )
{
buf[j] = src[0]*(1.f/255.f);
buf[j+1] = src[1]*(1.f/255.f);
@ -3227,7 +3267,23 @@ struct RGB2HLS_b
}
cvt(buf, buf, dn);
for( j = 0; j < dn*3; j += 3 )
j = 0;
#if CV_NEON
for ( ; j <= (dn - 8) * 3; j += 24)
{
float32x4x3_t v_src0 = vld3q_f32(buf + j), v_src1 = vld3q_f32(buf + j + 12);
uint8x8x3_t v_dst;
v_dst.val[0] = vqmovn_u16(vcombine_u16(vqmovn_u32(cv_vrndq_u32_f32(v_src0.val[0])),
vqmovn_u32(cv_vrndq_u32_f32(v_src1.val[0]))));
v_dst.val[1] = vqmovn_u16(vcombine_u16(vqmovn_u32(cv_vrndq_u32_f32(vmulq_f32(v_src0.val[1], v_scale))),
vqmovn_u32(cv_vrndq_u32_f32(vmulq_f32(v_src1.val[1], v_scale)))));
v_dst.val[2] = vqmovn_u16(vcombine_u16(vqmovn_u32(cv_vrndq_u32_f32(vmulq_f32(v_src0.val[2], v_scale))),
vqmovn_u32(cv_vrndq_u32_f32(vmulq_f32(v_src1.val[2], v_scale)))));
vst3_u8(dst + j, v_dst);
}
#endif
for( ; j < dn*3; j += 3 )
{
dst[j] = saturate_cast<uchar>(buf[j]);
dst[j+1] = saturate_cast<uchar>(buf[j+1]*255.f);
@ -3238,6 +3294,10 @@ struct RGB2HLS_b
int srccn;
RGB2HLS_f cvt;
#if CV_NEON
float32x4_t v_scale, v_scale_inv;
uint8x8_t v_alpha;
#endif
};