Otsu threshold is used to automatically perform clustering-based image thresholding or the reduction of a gray-level image to a binary image. The algorithm assumes that the image contains two classes of pixels following bi-modal histogram (foreground pixels and background pixels), it then calculates the optimum threshold separating the two classes.
Otsu method is used to find the threshold which can minimize the intra class variance which separates two classes defined by weighted sum of variances of two classes.
Where, w_1is the class probability computed from the histogram.
Otsu shows that minimizing the intra-class variance is the same as maximizing inter-class variance
Where,
is the class mean.
API Syntax
template<int SRC_T, int ROWS, int COLS,int NPC=1, int XFCVDEPTH_IN = _XFCVDEPTH_DEFAULT> OtsuThreshold(xf::cv::Mat<SRC_T, ROWS, COLS, NPC, XFCVDEPTH_IN> & _src_mat, uint8_t &_thresh)
Parameter Descriptions
The following table describes the template and the function parameters.
| Paramete r | Description |
|---|---|
| SRC_T | Input pixel type. Only 8-bit, unsigned, 1 channel is supported (XF_8UC1) |
| ROWS | Maximum height of input and output image. |
| COLS | Maximum width of input and output image (must be a multiple of 8, for 8-pixel operation) |
| NPC | Number of pixels to be processed per cycle; possible options are XF_NPPC1 and XF_NPPC8 for 1 pixel and 8 pixel operations respectively. |
| XFCVDEPTH_IN | Depth of the input image. |
| XFCVDEPTH_OUT | Depth of the output image. |
| _src_ma t | Input image |
| _thresh | Output threshold value after the computation |
Resource Utilization
The following table summarizes the resource utilization of the OtsuThreshold function, generated using Vivado HLS 2019.1 tool for the Xilinx Xczu9eg-ffvb1156-1-i-es1 FPGA, to process a grayscale HD (1080x1920) image.
| Operating Mode | Operating Frequency (MHz) | Utilization Estimate | ||||
|---|---|---|---|---|---|---|
| BRAM_18K | DSP_48Es | FF | LUT | CLB | ||
| 1 pixel | 300 | 8 | 49 | 2239 | 3353 | 653 |
| 8 pixel | 150 | 22 | 49 | 1106 | 3615 | 704 |
Performance Estimate
The following table summarizes the performance in different configurations, as generated using Vivado HLS 2019.1 tool for the Xilinx Xczu9eg-ffvb1156-1-i-es1, to process a grayscale HD (1080x1920) image.
| Operating Mode | Latency Estimate |
|---|---|
| Max Latency (ms) | |
| 1 pixel operation (300 MHz) | 6.92 |
| 8 pixel operation (150 MHz) | 1.76 |