Object grasping method, apparatus, computer device and storage medium

A computer device, comprising:

• a processor, a memory and a bus, the memory has machine-readable instructions executable by the processor stored thereon, the processor communicates with the memory through the bus when the computer device is running, and the machine-readable instructions, when executed by the processor, execute the steps of an object grasping method, comprising:
• in response to receiving a target object grasping instruction, determining a second confidence of each of candidate objects being a target object based on state description information of the target object comprised in the target object grasping instruction and a first confidence of each of the candidate objects in a current scene having each preset state;
• determining a target candidate object and a reward information after performing each preset action for the target candidate object based on the second confidence and the first confidence corresponding to each of the candidate objects, the preset action comprising grasping the target candidate object and asking a question;
• and
• determining a target preset action to be performed based on the reward information corresponding to each preset action, and performing the target preset action for the target candidate object,
• wherein the first confidence comprises a third confidence of each candidate object having each preset feature attribute, a fourth confidence of each candidate object being located in different preset position areas, and a fifth confidence of each candidate object having each preset position relationship with each of remaining candidate objects, and wherein the determining the second confidence of each of the candidate objects comprises:
• determining a sixth confidence of each of the candidate objects being the target object based on a target feature attribute comprised in the state description information of the target object and the third confidence of each of the candidate objects having the target feature attribute,
• determining a seventh confidence of the state description information of the target object comprising a target preset position area based on a second similarity between the state description information of the target object and standard position area description information corresponding to each preset position area, and determining an eighth confidence of each of the candidate objects being the target object based on the seventh confidence and the fourth confidence of each of the candidate objects being located in the target preset position area,
• determining a ninth confidence of the state description information of the target object comprising a target candidate object based on a third similarity between the state description information of the target object and standard candidate object description information corresponding to each of the candidate objects, determining a tenth confidence of the state description information of the target object comprising a target preset position relationship based on a fourth similarity between the state description information of the target object and standard position relationship description information corresponding to each preset position relationship, and
• determining an eleventh confidence of each of the candidate objects being the target object based on the ninth confidence, the tenth confidence and the fifth confidence of each of the candidate objects having the target preset position relationship with the rest each of the candidate objects, and
• determining the second confidence of each of the candidate objects being the target object based on the sixth confidence, the eighth confidence and the eleventh confidence.

A non-transitory computer-readable storage medium having a computer program stored thereon, the computer program, when run by a processor, executes the steps of an object grasping method, comprising:

• in response to receiving a target object grasping instruction, determining a second confidence of each of candidate objects being a target object based on state description information of the target object comprised in the target object grasping instruction and a first confidence of each of the candidate objects in a current scene having each preset state;
• determining a target candidate object and a reward information after performing each preset action for the target candidate object based on the second confidence and the first confidence corresponding to each of the candidate objects, the preset action comprising grasping the target candidate object and asking a question;
• and
• determining a target preset action to be performed based on the reward information corresponding to each preset action, and performing the target preset action for the target candidate object,
• wherein the first confidence comprises a third confidence of each candidate object having each preset feature attribute, a fourth confidence of each candidate object being located in different preset position areas, and a fifth confidence of each candidate object having each preset position relationship with each of remaining candidate objects, and wherein the determining the second confidence of each of the candidate objects comprises:
• determining a sixth confidence of each of the candidate objects being the target object based on a target feature attribute comprised in the state description information of the target object and the third confidence of each of the candidate objects having the target feature attribute,
• determining a seventh confidence of the state description information of the target object comprising a target preset position area based on a second similarity between the state description information of the target object and standard position area description information corresponding to each preset position area, and determining an eighth confidence of each of the candidate objects being the target object based on the seventh confidence and the fourth confidence of each of the candidate objects being located in the target preset position area,
• determining a ninth confidence of the state description information of the target object comprising a target candidate object based on a third similarity between the state description information of the target object and standard candidate object description information corresponding to each of the candidate objects, determining a tenth confidence of the state description information of the target object comprising a target preset position relationship based on a fourth similarity between the state description information of the target object and standard position relationship description information corresponding to each preset position relationship, and
• determining an eleventh confidence of each of the candidate objects being the target object based on the ninth confidence, the tenth confidence and the fifth confidence of each of the candidate objects having the target preset position relationship with the rest each of the candidate objects, and
• determining the second confidence of each of the candidate objects being the target object based on the sixth confidence, the eighth confidence and the eleventh confidence.