 |
PMSIS API
|
 |
|
PMSIS API
|
|
Data Structures | |
| struct | pi_cpi_conf |
| CPI device configuration structure. More... | |
Enumerations | |
| enum | pi_cpi_format_e |
| Image format identifier. More... | |
Functions | |
| void | pi_cpi_conf_init (struct pi_cpi_conf *conf) |
| Initialize a CPI configuration with default values. More... | |
| int | pi_cpi_open (struct pi_device *device) |
| Open a CPI device. More... | |
| void | pi_cpi_close (struct pi_device *device) |
| Close an opened CPI device. More... | |
| void | pi_cpi_capture (struct pi_device *device, void *buffer, int32_t bufferlen) |
| Capture a sequence of samples. More... | |
| void | pi_cpi_capture_async (struct pi_device *device, void *buffer, int32_t bufferlen, pi_task_t *task) |
| Capture a sequence of samples asynchronously. More... | |
| static void | pi_cpi_control_start (struct pi_device *device) |
| Start capturing samples. More... | |
| static void | pi_cpi_control_stop (struct pi_device *device) |
| Stop capturing samples. More... | |
| static void | pi_cpi_set_format (struct pi_device *device, pi_cpi_format_e format) |
| Set frame format. More... | |
| static void | pi_cpi_set_rowlen (struct pi_device *device, uint16_t rowlen) |
| Set frame row length. More... | |
| static void | pi_cpi_set_frame_drop (struct pi_device *device, uint32_t nb_frame_dropped) |
| Configure frame drop. More... | |
| static void | pi_cpi_set_filter (struct pi_device *device, uint32_t r_coeff, uint32_t g_coeff, uint32_t b_coeff, uint32_t shift) |
| Configure frame filtering. (ONLY available in GAP8) More... | |
| static void | pi_cpi_set_rgb_sequence (struct pi_device *device, uint8_t rgb_seq) |
| Configure RGB Sequence. More... | |
| static void | pi_cpi_set_slice (struct pi_device *device, uint32_t x, uint32_t y, uint32_t w, uint32_t h) |
| Configure frame slicing. More... | |
| static void | pi_cpi_set_sync_polarity (struct pi_device *device, uint8_t vsync_pol_ena, uint8_t hsync_pol_ena) |
| Configure vsync/hsync polarity. More... | |
This API provides support for capturing images from an image sensor through the Camera Parallel Interface (CPI) and processing them.
| enum pi_cpi_format_e |
This can be used to describe the format of the image going through the interface.
| void pi_cpi_capture | ( | struct pi_device * | device, |
| void * | buffer, | ||
| int32_t | bufferlen | ||
| ) |
Queue a buffer that will receive samples from the CPI interface. This function is synchronous and will block the caller until the specified amount of bytes is received. The buffer will receive samples only if the interface is started. Depending on the chip, there may be some restrictions on the memory which can be used. Check the chip-specific documentation for more details.
| device | A pointer to the structure describing the device. |
| buffer | The memory buffer where the captured samples will be transferred. |
| bufferlen | The size in bytes of the memory buffer. |
| void pi_cpi_capture_async | ( | struct pi_device * | device, |
| void * | buffer, | ||
| int32_t | bufferlen, | ||
| pi_task_t * | task | ||
| ) |
Queue a buffer that will receive samples from the CPI interface. This function is asynchronous and will not block the caller. It is possible to call it several times in order to queue several buffers. At a minimum 2 buffers should be queued to ensure that no data sampled is lost. This is also the most efficient way to retrieve data from the CPI device. You should always make sure that at least 2 buffers are always queued, by queuing a new one as soon as the current one is full. The buffer will receive samples only if the interface is started. A task must be specified in order to specify how the caller should be notified when the transfer is finished. Depending on the chip, there may be some restrictions on the memory which can be used. Check the chip-specific documentation for more details.
| device | A pointer to the structure describing the device. |
| buffer | The memory buffer where the captured samples will be transferred. |
| bufferlen | The size in bytes of the memory buffer. |
| task | The task used to notify the end of transfer. See the documentation of pi_task for more details. |
| void pi_cpi_close | ( | struct pi_device * | device | ) |
This function can be called to close an opened CPI device once it is not needed anymore in order to free all allocated resources. Once this function is called, the device is not accessible anymore and must be opened again before being used.
| device | A pointer to the structure describing the device. |
| void pi_cpi_conf_init | ( | struct pi_cpi_conf * | conf | ) |
The structure containing the configuration must be kept alive until the camera device is opened.
| conf | A pointer to the CPI configuration. |
|
inlinestatic |
The samples arriving at the CPI interface are dropped until the interface is started by calling this function. Once started, the sampling will start at the next beginning of frame.
| device | A pointer to the structure describing the device. |
|
inlinestatic |
The samples arriving at the CPI interface are dropped after this call.
| device | A pointer to the structure describing the device. |
| int pi_cpi_open | ( | struct pi_device * | device | ) |
This function must be called before the CPI device can be used. It configures the specified device that can then be used to refer to the opened device when calling other functions.
| device | A pointer to the structure describing the device. This structure muts be allocated by the caller and kept alive until the device is closed. |
|
inlinestatic |
Configure how to filter the input pixels to produce the 8bits output pixels. Each channel is multiplied by a coefficient. They are then summed together and shifted right to obtain an 8 bit pixel.
| device | A pointer to the structure describing the device. |
| r_coeff | Red channel is multipled by this coefficient. |
| g_coeff | Green channel is multipled by this coefficient. |
| b_coeff | Blue channel is multipled by this coefficient. |
| shift | The sum of all channels multiplied by their coefficient is shifted right by this value to fit an 8 bit pixel. |
|
inlinestatic |
This can be used either when filtering is not active to specify which part of the samples to keep, or when filtering is active, to specify what is the input format, to be able to properly convert the image.
| device | A pointer to the structure describing the device. |
| format | The format of the input frame. |
|
inlinestatic |
For each sampled frame, the specified number of frames will be dropped before the next frame is sampled.
| device | A pointer to the structure describing the device. |
| nb_frame_dropped | Number of frames to drop. Can be set to 0 to deactivate this feature. |
|
inlinestatic |
Configure how the RGB be saved to L2 memory. (ONLY available in GAP9)
| device | A pointer to the structure describing the device. |
| rgb_seq | Sequence of RGB: 3'h0 - RGB, 3'h1 - RBG, 3'h2 - GRB, 3'h3 - GBR, 3'h4 - BRG, 3'h5 - BGR, |
|
inlinestatic |
Set the row length of each frame, should be the same as the camera output. Will be set at the beginning of cpi open.
| device | A pointer to the structure describing the device. |
| rowlen | The rowlen of each frame, should be divide by the datasize of CPI. For example, if the picture size is 320*240, cpi transfer datasize is 16bits, the rowlen should be: 320/2 (bytes). |
|
inlinestatic |
Configure how to extract a window slice out of the input frame.
| device | A pointer to the structure describing the device. |
| x | x position of the window to extract. |
| y | y position of the window to extract. |
| w | Width of the window to extract. |
| h | Height of the window to extract. |
|
inlinestatic |
Configure to reverse the vsync and/or hsync polarity. (ONLY available in GAP9)
| device | A pointer to the structure describing the device. |
| vsync_pol_ena | Enable the vsync polarity reversal. |
| hsync_pol_ena | Enable the hsync polarity reversal. |
