GrayScaleObservation

class GrayScaleObservation(env, keep_dim=True, key='img')[source]

Bases: ObservationWrapper

Convert the image observation from RGB to gray scale.

Mostly the same as the open ai gym implementation except this allows for the wrapper to be applied to a single element inside a Dict observation.

__init__(env, keep_dim=True, key='img')[source]

Initialze an object.

Parameters:

env (gym environment) – Environment to wrap.
keep_dim (bool, optional) – Flag indicating if the channel dimension should be kept. The default is True.
key (string, optional) – Key in teh dictionary corresponding to the image. The default is ‘img’.

Methods

`__init__`	Initialze an object.
`class_name`	Returns the class name of the wrapper.
`close`	Closes the environment.
`observation`	Generates the proper observation.
`render`
`reset`	Resets the environment, returning a modified observation using `self.observation()`.
`seed`	Seeds the environment.
`step`	Returns a modified observation using `self.observation()` after calling `env.step()`.

Attributes

`action_space`	Returns the action space of the environment.
`metadata`	Returns the environment metadata.
`np_random`	Returns the environment np_random.
`observation_space`	Returns the observation space of the environment.
`render_mode`	Returns the environment render_mode.
`reward_range`	Return the reward range of the environment.
`spec`	Returns the environment specification.
`unwrapped`	Returns the base environment of the wrapper.

reset(**kwargs): Resets the environment, returning a modified observation using self.observation().

step(action): Returns a modified observation using self.observation() after calling env.step().

property action_space: Space[ActType]: Returns the action space of the environment.

property np_random: RandomNumberGenerator: Returns the environment np_random.

property observation_space: Space: Returns the observation space of the environment.

property reward_range: Tuple[SupportsFloat, SupportsFloat]: Return the reward range of the environment.