Read images from sd card

#1

Hello guys, I’m currently trying to run a image classification in an arduino nano 33 ble sense, the problem is that it works perfectly with the ov7675 example but i don’t know how to make it work reading a set of images from a sd card. The images are in .bmp format and (what I understood) I need a rgb888 raw pixel buffer first, then do the resize and crop (if needed) and then make a callback function returning that buffer for the signal struct to run the inference. The problem is I don’t know how to convert from a BMP buffer to what i need. It is so easy with the esp32-camera library because it has conversions… but with arduino is not as easy

Thanks you a lot

#2

Hi @josemimo2

In the Arduino library that you generate from your Edge Impulse project there are a number of examples:
File> Examples>Examples from custom libaries> project_name_camera

will have an example that you can refer to for the streaming and conversion you are looking for:

/**
 * @brief   Setup image sensor & start streaming
 *
 * @retval  false if initialisation failed
 */
bool ei_camera_init(void) {
    if (is_initialised) return true;
    
    if (!Cam.begin(QQVGA, RGB565, 1)) { // VGA downsampled to QQVGA (OV7675)
        ei_printf("ERR: Failed to initialize camera\r\n");
        return false;
    }
    is_initialised = true;

    return true;
}

/**
 * @brief      Convert RGB565 raw camera buffer to RGB888
 *
 * @param[in]   offset       pixel offset of raw buffer
 * @param[in]   length       number of pixels to convert
 * @param[out]  out_buf      pointer to store output image
 */
int ei_camera_cutout_get_data(size_t offset, size_t length, float *out_ptr) {
    size_t pixel_ix = offset * 2; 
    size_t bytes_left = length;
    size_t out_ptr_ix = 0;

    // read byte for byte
    while (bytes_left != 0) {
        // grab the value and convert to r/g/b
        uint16_t pixel = (ei_camera_capture_out[pixel_ix] << 8) | ei_camera_capture_out[pixel_ix+1];
        uint8_t r, g, b;
        r = ((pixel >> 11) & 0x1f) << 3;
        g = ((pixel >> 5) & 0x3f) << 2;
        b = (pixel & 0x1f) << 3;

        // then convert to out_ptr format
        float pixel_f = (r << 16) + (g << 8) + b;
        out_ptr[out_ptr_ix] = pixel_f;

        // and go to the next pixel
        out_ptr_ix++;
        pixel_ix+=2;
        bytes_left--;
    }

    // and done!
    return 0;
}

There are also a number of questions on the Arduino forum regarding this that may help:

Best

Eoin