OpenCV + gym-retro：输入图像中的通道数无效

Question

一直在摆弄 gym-retro 和 OpenCV。我一直在查看其他代码示例和教程。其中一些似乎以相同的方式编码，但是当我这样做时，出现以下错误。是否有某种类型的更新或什么？欢迎任何关于修复的建议。我可以注释掉重塑和转换为灰度并且它有效。然而，我向我的 NN 提供了太多信息。

import retro
import numpy as np
import cv2
import neat
import pickle

env = retro.make('SuperMarioBros3-Nes', '1Player.World1.Level1')

def eval_genomes(genomes, config):
    for genome_id,genome in genomes:
        ob = env.reset()
        ac = env.action_space.sample()
        inx, iny, inc = env.observation_space.shape
        inx = int(inx/8)
        iny = int(iny/8)
        net = neat.nn.recurrent.RecurrentNetwork.create(genome, config)
        current_max_fitness = 0
        fitness_current = 0
        frame = 0
        counter = 0
        xpos = 0
        xpos_max = 0
        done = False
        

        while not done:
            env.render()
            frame += 1
            #print(ob)
            
            ob = cv2.resize(ob, (inx,iny))
            ob = cv2.cvtColor(ob, cv2.COLOR_BGR2GRAY)
            ob = np.reshape(ob, (inx,iny))
            imgarray = np.ndarray.flatten(ob)
        
        nnOutput = net.activate(imgarray)
        
        ob, rew, done, info = env.step(nnOutput)
        #imgarray.clear()

config = neat.Config(neat.DefaultGenome, neat.DefaultReproduction,neat.DefaultSpeciesSet,neat.DefaultStagnation,'config-feedforward')

p = neat.Population(config)

winner = p.run(eval_genomes)

(gameai) C:\Users\dgilk\anaconda3\envs\gameai>python mario.py
Traceback (most recent call last):
  File "mario.py", line 45, in <module>
    winner = p.run(eval_genomes)
  File "C:\Users\dgilk\anaconda3\envs\gameai\lib\site-packages\neat\population.py", line 89, in run
    fitness_function(list(iteritems(self.population)), self.config)
  File "mario.py", line 32, in eval_genomes
    ob = cv2.cvtColor(ob, cv2.COLOR_BGR2GRAY)
cv2.error: OpenCV(4.4.0) c:\users\appveyor\appdata\local\temp\pip-req-build-k8sx3e60\opencv\modules\imgproc\src\color.simd_helpers.hpp:92: error: (-2:Unspecified error) in function '__cdecl cv::impl::`anonymous-namespace'::CvtHelper<struct cv::impl::`anonymous namespace'::Set<3,4,-1>,struct cv::impl::A0xbf2c9cd3::Set<1,-1,-1>,struct cv::impl::A0xbf2c9cd3::Set<0,2,5>,2>::CvtHelper(const class cv::_InputArray &,const class cv::_OutputArray &,int)'
> Invalid number of channels in input image:
>     'VScn::contains(scn)'
> where
>     'scn' is 1

更新：这是缩小图像的输出。好像有颜色。 snippet of observation window

Answer 1

您正在此处创建形状为 (inx,iny) 的 cv2 对象

   ob = cv2.resize(ob, (inx,iny)) # 1
   ob = cv2.cvtColor(ob, cv2.COLOR_BGR2GRAY) # 2

cv2.COLOR_BGR2GRAY 需要一个具有形状 (inx, iny, 3) 的彩色图像，因此请检查您需要 'ob' 的形状，第 2 行需要什么对于？

OpenCV + gym-retro：输入图像中的通道数无效

OpenCV + gym-retro: Invalid number of channels in input image

python

python-3.x

opencv

machine-learning

openai-gym