176 rga_buffer_t src;
178 memset(&src, 0, sizeof(rga_buffer_t));
179
180 // Source is NV12 (YUV420SP)
181 src.vir_addr = vi_data;
182 src.phy_addr = NULL; // Not using physical address
183 src.fd = -1; // No dma_buf
184 src.width = width;
185 src.height = height;
186 src.wstride = width;
187 src.hstride = height;
188 src.format = RK_FORMAT_YCbCr_420_SP;
189 src.color_space_mode = 0; // Default color space
190 src.global_alpha = 0xFF;
191 src.rd_mode = 0;
192
193 // Destination is BGR888
194 rga_buffer_t dst;
195 memset(&dst, 0, sizeof(rga_buffer_t));
196
197 dst.vir_addr = bgr.data;
198 dst.phy_addr = NULL;
199 dst.fd = -1;
200 dst.width = width;
201 dst.height = height;
202 dst.wstride = width;
203 dst.hstride = height;
204 dst.format = RK_FORMAT_BGR_888;
205 dst.color_space_mode = 0;
206 dst.global_alpha = 0xFF;
207 dst.rd_mode = 0;
208
209 //color conversion
210 int ret = imcvtcolor(src, dst, src.format, dst.format);
211 if (ret != IM_STATUS_SUCCESS) {
212 printf("RGA imcvtcolor failed: %d\n", ret);
213 }
why this code is not working?
what i'm missing?
Can anyone please help with this issue.
RGA in luckfox pico ultra w
Hello, we do not offer technical support for the secondary development of the software. For more details, you can refer to https://github.com/airockchip/librga. The example routines provided by Rockchip include the standard usage of imcvtcolor.
Is the luckfox pico sdk, opencv library supports face detection?
The luckfox pico will consume excessive memory resources when opening OpenCV and the performance is not good. Generally, the lightweight and optimized OpenCV-Mobile is used. There is an introduction on how to implement face detection using OpenCV-Mobile + RKNN on the wiki.
I'm using the Opencv which is in rkmpi example of rtsp retinaface. The avg use of CPU is 60-70% and RAM around 23%(MAX).
I have removed the rtsp as i'm using on board display.
Is this much CPU and RAM use normal?
I have removed the rtsp as i'm using on board display.
Is this much CPU and RAM use normal?
Running "luckfox_pico_rtsp_retinaface" results in CPU usage around 60-70% and RAM usage around 20-30%, which are all within normal ranges.
Ok
thank you
thank you